SYLLLIGEUSS ae. NATIONAL MUSEUM OF NATURAL SCIENCES SESE 55555555555 TYAN Ee SESS 5555555555 No. 33 MUSEE NATIONAL DES SCIENCES NATURELLES C. R. Harington, editor CLIMATIC CHANGE IN CANADA 2 MUSEES NATIONAUX DU CANADA OTTAWA SYLLOGEUS is a publication of the National Museum of Natural Sciences, National Museums of Canada, designed to permit the rapid dissemination of information pertaining to those disciplines and educational functions for which the National Museum of Natural Sciences is responsible. In the interests of making information available quickly, normal publishing procedures have been abbreviated. Articles are published in English, in French, or in both languages, and the issues appear at irregular intervals. A complete list of the titles issued since the beginning of the series (1972) and individual copies of this number are available by mail from the National Museum of Natural Sciences, Ottawa, Canada. KlA OM8 La collection SYLLOGEUS, publiée par le Musée national des sciences naturelles, Musée nationaux du Canada, a pour but de diffuser rapidement le résultat des travaux dans les domaines scientifique et éducatif qui sont sous la direction du Musée national des sciences naturelles. Pour assurer la prompte distribution de cette publication, on a abregé les étapes de la rédaction. Les articles sont publiés en francais, en anglais ou dans les deux langues, et ils paraissent irréguliérement. On peut obtenir par commande postale la liste des titres de tous les articles publiés depuis le début de la collection (1972) et des copies individuelles de ce numéro, au Musée national des sciences naturelles, Ottawa, Canada. K1A OM8 Syllogeus Series No. 33 Série Syllogeus No. 33 (c) National Museums of Canada 1981 (c) Musée nationaux du Canada 1981 Printed in Canada Imprimé au Canada ISSN 0704-576X CLIMATIC - CHANGE IN CANADA : 2 Saene sen ina ea | NORMES. LÉ Sees Be Vee»... ERE SS a 2 See eee ew... Digitized by the Internet Archive in 2011 with funding from California Academy of Sciences Library htip://www.archive.org/details/syllogeus33nati CLIMATIC CHANGE IN CANADA - 2 National Museum of Natural Sciences Project on Climatic Change in Canada During the Past 20,000 Years Edited by C.R. Harington Syllogeus No. 33 National Museums of Canada Les Musées nationaux du Canada National Museum of Natural Sciences Musée national des Sciences naturelles ACKNOWLEDGEMENTS Members of the NMNS climatic change project are grateful to: Dr. Louis Lemieux, (Director, National Museum of Natural Sciences), Mr. F. Hugh Schultz (Assistant Director, Research and Operations, National Museum of Natural Sciences) and Dr. Dale A. Russell (Chief, Paleobiology Division) for their continuing support; Mrs. Heather Shannon (Executive . Assistant to the Director) and Mrs. Helen Davies (Special Projects Officer, National Museum of Natural Sciences) for help in arranging meetings and luncheons; Mrs. J. McConnell (Head, Word Processing Centre, National Museums of Canada) and Ms. Barbara Renaud (Systemhouse Ltd.) for their help in typing the manuscripts for publication. The editor wishes to thank Mrs. Gail Rice (Secretary, Paleobiology Division) for assistance in all phases of the project and for help in preparation of the final draft of this publication. It is with great regret that the death of Gordon Manley, the outstanding English paleoclimatologist and geographer, is noted. The editor is grateful to The Royal Geographical Society for permission to reprint his obituary at the end of this volume. CONTENTS Preface G.D. Hobson Introduction C.R. Harington There's a Chill in the Air J. Gordon Ogden, III Recent Changes in Temperature in Canada, and comments on future climatic change M.O. Berry Analysis of Recent Climatic Changes in Québec: Some Preliminary Data Claude Hillatre-Marcel, Serge Oechtetti, Louise Marchand and Reiner Rajewtez Analysis of Historical Evidence of Climatic Change in Western and Northern Canada A.J.W. Catehpole and T.F. Ball Review of Dendroclimatology in the Forest-Tundra Ecotone of Alaska and Canada Gordon C. Jacoby and Linda D. Ulan Dendrochronological Studies on the Coasts of James Bay and Hudson Bay (Parts 1 and 2) M.L. Parker, L.A. Jozsa, Sandra G. Johnson and Paul A. Bramhall Impact of Climatic Variation on Boreal Forest Biomass Production John M. Powell Studying Climatic Change from Canadian High Arctic Ice Cores R.M. Koerner and D.A. Fisher Obituary: Gordon Manley (1902-1980) 10 19 28 48 97 129 189 195 DE) PREFACE Are we on a warming trend or a cooling trend? Depending upon whom you ask, the answer may be either one or the other. In fact, there is so much controversy over this question that the simplest answer may be, "Yes, we are on a warming or a cooling trend." Why did Parry get as far west as Winter Harbour in 1818 while Franklin met disaster in 1845/47? Was it entirely due to ice conditions? Or were there differences in climate? Long-term data in climatology are scarce and we therefore must go into the records retained in sediment samples, tree rings, ice cores, etc. We know now that the Climatic Optimum can be recognized and that volcanic activity can be dated. In recent years there have been significant advances in the interpretation of data encased in cores of various types. However, a more accurate prediction of future climate is required for general planning purposes in this country. In the field of agriculture new strains of grain would have to be developed if we were facing continued cooling trends. The intensive use of the Northwest Passage and the Great Lakes System for marine transportation of natural and manufactured resources is very dependent upon the climate. Our water resources and food growth are dependent upon precipitation which can be dependent upon warm or cold climate. Climate changes would effect population migrations and thus affect the social and cultural framework of North America. Perhaps the most important impact of climate upon Canada itself would be upon our dependence on hydrocarbon resources. If we are on a warming trend our need for an alternate source of fuel is not quite as critical as if we are on a cooling trend. Perhaps climate should be one of the unknown terms in the National Energy Policy equation. In short, we must pursue the study of the past so that we may take an educated guess at the future. In the very near future we must answer the question as to whether we are indeed on a cooling curve or a warming curve or whether we are merely on a small bump on either one of them. There are so many factors to consider and, unfortunately, we are not conducting enough scientific research into the problem to answer some very practical and simple questions. G.D. Hobson, Director Polar Continental Shelf Project INTRODUCTION C.R. Harington* The reasons for initiation of the National Museum of Natural Sciences climatic change project, its past history, goals and approaches have been described elsewhere (Harington 1980). This is the second publication arising from the project, the first being Climatte Change in Canada (Syllogeus No. 26, 1980). Whereas the latter volume was the product of a pilot study and included mainly summaries of information and methods of studying various types of proxy data, this volume covers a broader field and provides more in the way of results and their interpretation. We are particularly fortunate to have contributions from associates of the project (M.O. Berry (Atmospheric Environment Service), Gordon C. Jacoby and Linda D. Ulan (Lamont-Doherty Geological Observatory), John M. Powell (Canadian Forestry Service), and R.M. Koerner and David Fisher (Polar Continental Shelf Project)), in addition to those from members, whose work is at least partly funded through the climatic change project. Papers in this volume have been organized in approximate sequence ranging from analyses of relatively recent meteorological records to glaciological research covering a much longer period of geological time. Thus, they extend back from studies of meteorological records (e.g. Ogden, Berry, Hillaire-Marcel et al.), to those dealing with climatic data derived from Hudson's Bay Company records covering the last 200 or more years (Catchpole and Ball), to those involving tree-ring studies reaching back more than 500 years (Jacoby and Ulan, Parker et al., Powell), to glaciological studies capable of yielding paleoclimatic data for the last 100,000 or more years. Some highlights of these papers are reviewed below in order to: (a) give a brief idea of their scope; (b) show where certain studies complement one another, and, (c) underline similarities and differences, where climatic predictions are discussed. Ogden stresses the socio-economic effects of climatic change in Nova Scotia. He notes that a few tenths of a degree increase in coastal water temperatures results in reduced lobster catches. During the winter of 1971-72, unusual coldness led to an additional *Paleobiology Division, National Museum of Natural Sciences, National Museums of Canada, Ottawa, Ontario, K1A OM8 consumption of some 15 million gallons (68 million litres) of fuel oil in the Halifax- Dartmouth region alone. The cost, at that time, amounted to more than 6 million dollars. He also observes that the loss of green peppers as a cash crop in Nova Scotia reflects the general decrease in growing degree-days since the early 1950s. Ogden advocates a regional multidisciplinary effort to produce contingency plans in the fields of agriculture, forestry and energy in anticipation of future climatic trends. From his study of meteorological records, Berry shows that the eastern half of Canada was colder by 1969-1978 than in the decade beginning 1949, with the cooling trend in most areas (the Lower Great Lakes, Upper Saint Lawrence and Maritimes excepted) intensifying from the second to last decade of that period. In contrast, much of the western half of the country was warmer by the third decade (except for cooling in northern British Columbia and, part of the Yukon Territory). So, since 1949 at least, it is important to recognize that differences occurred in regional temperature trends for Canada. Although many problems are involved in attempting to predict climate far into the future, Berry foresees a long-term downward trend in global temperatures, with successively more severe cooling peaks at 5, 23 and 59 millenia from now. Superimposed on such a trend would be shorter term variations, such as a possible increase by 2-3°C in global mean temperature (with greater warming at higher latitudes) during the next 50 to 100 years. The latter warming could result from increasing carbon dioxide concentration in the atmosphere due mainly to fossil-fuel consumption. Hillaire-Marcel et al. compare instrumental temperature records for Montréal and Toronto since 1840. Interestingly, over this period, they find little difference in summer temperature between these two cities, despite their latitudinal difference. A moderate cooling of climate is predicted over the next 50 years, if the temperature maximum observed in the 1950s is used as a basis. They point out, cogently, that the regular decline in average temperatures since the 1950s (despite increasing injections of carbon dioxide into the atmosphere by human activity) indicates that natural control of climate overrides the human impact. Catchpole and Ball report on paleoclimatic data they derive from studies of Hudson's Bay Company records including: dates of freezing and breaking of ice in James Bay estuaries; sea-ice conditions in Hudson Strait as determined by information in Hudson's Bay Company ships' logs; and the development of a computer-coding system for retrieval and analysis of climatic information from post journals. Interpretation of these and subsequent findings in the light of changing climate will be presented later. I wish to emphasize the great amount of time and care it takes to derive useful information from such historic records. For example, coding of transcribed climatic data from the York Factory journals, and direct coding of the Churchill data took Tim Ball approximately 2.5 years! Undoubtedly, computerization of climatic data in the Hudson's Bay Company records is the most efficient way to handle it -- it saves researchers going through the documents again and again for each new category of information. These records, when finally mined of their vast riches, will provide an extremely well-rounded picture of climate in central and western Canada during the past 200 years. As the Prairies are neglected in this publication, it is worth noting that L.V. Hills (a member of the NMNS climatic change project) is studying the history of the Peace River grasslands. Further, Currie and Venkatarangan (1978, p. 22, Figure 11) are convinced that solar disturbances, as indicated by relative sunspot numbers, contribute to a variation of Prairie precipitation — above normal precipitation occurring in several years centred on the sunspot minimum. Jacoby and Ulan state that a climatically sensitive tree-ring series longer than 500 years can be developed in the forest-tundra ecotone of Canada and Alaska. Both temperature and moisture information are recorded in the tree rings. Their chronology from the Coppermine River area, Northwest Territories, will be awaited with interest because it extends back to 1428 and is particularly sensitive climatically. The effects of cooling during the Little Ice Age, and recent Northern Hemisphere warming (to about 1950) appear in tree-ring records from several of their sites. Presumably their, and Parker's, work on sites near the west coast of Hudson Bay will yield results that may be checked against paleoclimatic data derived from Hudson's Bay Company journals from the same region by Catchpole and Ball. In 1979, Parker and Jozsa collected samples from four sites along the eastern coasts of James Bay and Hudson Bay in order to build tree-ring width and density chronologies for that region. Their specific goal was to provide climatic data from this source for comparison with early instrumental and documentary climatic information derived by Cynthia Wilson from the Great Whale River journals of the Hudson's Bay Company. Because they are most pertinent to Wilson's study, only the Cri Lake tree-ring data (extending back to 1732) are reported here. Tree-ring data from other sites (Matagami, Radisson and Richmond Gulf) sampled in 1979 will be processed later. Powell notes that two north-south dendrochronological transects through the Boreal Forest in western Canada will tie in with studies by Jacoby and others near the northern tree line. This work, undertaken by Parker and his colleagues for the Canadian Forestry Service, will provide a network of tree-ring sites that should lead to more comprehensive paleoclimatic analyses of large areas of central Canada. One of the main goals of Powell's project is to determine the impact of climatic change upon tree growth in the Boreal Forest. This program exemplifies the distinct economic value implicit in much paleoclimatic research. From a study of Canadian High Arctic ice cores, Koerner and Fisher conclude that the large ice caps on Devon, Ellesmere and Axel Heiberg Islands are about 100,000 or more years old, and that the Meighen Ice Cap and Ward-Hunt Ice Shelf probably began to grow 4,000 years ago. Ice caps, such as that on Devon Island, have yielded valuable temperature information covering the past 10,000 years. From such data, Koerner and Fisher find that the High Arctic climate has been cooling over the past 5,000 years, and they predict (disregarding a possible warming due to carbon dioxide increase) a continuation of the cooling trend to 1990 and perhaps beyond. They also indicate that shorter snow-free seasons and less navigable sea-ice conditions can be expected — points of considerable interest to those concerned with shipping in the region. Although it is difficult to fully understand the carbon cycle (Hare 1980), there is little reason to doubt that rising carbon dioxide concentrations (mainly as a result of people burning fossil fuels) will increase atmospheric temperature in the near future. A major question arising from several discussions in this volume is, "Will the warming override what appears to be a natural cooling trend?" No one can say. If it does, and the warming is strong enough, I suggest that Canada would not necessarily be adversely affected. Northern limits of crops, and the Boreal Forest would shift northward, and passages between the Canadian Arctic Islands would become open to shipping and perhaps richer in some forms of marine life. Adaptability, based on a sound knowledge of climatic change in Canada during the past 20,000 years, will be of key importance in facing such problems. In conclusion, I wish to mention a recent meeting of the NMNS climatic change project and a meeting being planned for 1982. On December 1, 1980, the third meeting of this group took place in Ottawa. Twenty people attended, and twelve papers were presented by scientists from Vancouver, Edmonton, Winnipeg, Ottawa, Downsyiew, Montréal and Halifax. The project is at a most interesting and important juncture. Soon we will be able to test the value of various types of proxy data (e.g. climatic records from Hudson's Bay Company documents can be checked against sensitive tree-ring records: both cover approximately the last 300 years in the southern Hudson Bay - James Bay region). Further, it is now possible to recreate ancient surface and upper air pressure patterns (which are important in elucidating causes of climatic change) for the same region by computer analysis of wind directions recorded in Hudson's Bay Company post journals. Although there will be no meeting of project members and associates in 1981, a fourth meeting is being planned for 1982, when discussions will focus on "Critical Periods in the Climate of Canada Over the Past 20,000 years." Perhaps by concentrating on such events (e.g. the postglacial melting about 10,000 years ago, the Little Climatic Optimum about 1000 A.D., the Little Ice Age about 1500-1850 A.D., times of major volcanic eruptions and years of extreme cold, drought, etc.), we could clarify their nature, causes and impacts as viewed from various fields. If these catastrophic points can be established, then it will be easier to fill in the more uniform periods, and ultimately to determine if cycles are discernible. Prediction is, of course, an important goal in this type of research, but regardless of that, the results should be of great interest and value academically (e.g. to biogeographers, paleoecologists, archaeologists and historians). REFERENCES Hare, F.K. 1980. The planetary environment: fragile or sturdy? Geogr. J. 146, Pt. 3:379-395. Harington, C.R. 1980. The impact of changing climate on people in Canada; and the National Museum of Natural Sciences Climatic Change Project. In: Climatic Change in Canada. Edited by: C.R. Harington. Syllogeus No. 26. National Museums of Canada. pp. 5-15. Currie, B.W., and P. Venkatarangan. 1978. Relationships between solar disturbances and the precipitation and temperature on the Canadian Prairies. Inst. Space and Atmospheric Stud., Univ. Saskatchewan (Saskatoon) Rept. 58 pp. THERE'S A CHILL IN THE AIR J. Gordon Ogden, III* Climate is usually regarded as a nuisance we endure while waiting for the weather to get better. Far from being stable, Nova Scotia's climate has varied dramatically in the past and may be expected to do so in the future. Less than 15,000 years ago all of the Maritimes region was deeply buried beneath glacier ice, sea levels were several hundred feet lower, and shorelines were several tens of miles seaward of their present positions. Climatic changes caused the melting and retreat of the glaciers resulting in ameliorating temperatures, rising sea levels, and rebound of the earth's crust as it was released from the crushing weight of glacier ice. By 10,500 years ago, the first Nova Scotians apparently passed through customs and arrived at Debert near Truro, Nova Scotia (MacDonald 1968; Stuckenrath 1966; Terasmae 1972). Since that time Nova Scotia has experienced much milder climates than at present, as well as some cooler periods. More recently, the interval known as the Little Climatic Optimum of about 1000 AD permitted, if not encouraged, westward exploration and colonization by the Vikings. Similarly, the climatic deterioration beginning in the 12th century AD not only sent the Vikings scurrying home and wiped out their settlements in North America and Greenland, but also culminated in the Little Ice Age of the 18th century, during which farms and villages were overwhelmed by advancing valley glaciers in alpine regions of western Europe. Ladurie (1971) records the devestation of entire villages and also mentions the effects of deteriorating climate upon wine harvests and agricultural production in the 18th and 19th centuries. An example of the dramatic relationship between the Little Ice Age and climatic change can be found in Irish history. Following introduction of the potato to Ireland in the 1820s, newly found prosperity allowed a virtual doubling of the population by 1840. But, beginning in 1846 and continuing for nearly a decade, a series of long cold winters, late springs, and * Department of Biology, Dalhousie University, Halifax, Nova Scotia, B3H 4J1 10 cold foggy summers encouraged widespread growth of the potato blight fungus, resulting in at least six crop failures in the 10-year period. Starvation and emigration not only decimated the population but dramatically changed the life style of the survivors. Even today, marriage is often delayed and many families are childless resulting in decreasing population. Another result of this climatic tragedy is reflected in the country's music. All of the gay and lively jigs, reels and kerry dances antedate the blight, and the dirges and laments reflect the hardship partly due to the famine times. The climatic change associated with this disaster was a decline of less than 2°C in annual temperature. Subtle changes in climatic parameters such as temperature, precipitation, or air-mass circulation can have broad biological and economic effects. Spruce budworms flourish when conditions are warm and dry in June, while a few tenths of a degree increase in coastal water temperatures result in reduced lobster landings. A 0.5°C drop in average winter temperature (November to March) results in use of approximately an extra 10 gallons (46 litres) of heating oil per household in Canada. The winter of 1971-72 resulted in consumption of an additional 15 million gallons (68 million litres) of fuel oil in the Halifax-Dartmouth region alone, attributed to the weather. According to my latest fuel bill, that's more than 6 million dollars. Energy and economic policies which ignore the possibilities of a continuation of climatic deterioration since 1940 threaten a serious energy shortfall which may have important economic effects. Man's energy requirements are now a distinct geological and climatic force. Increasing levels of carbon dioxide from fuel combustion may increase temperatures at the earth's surface due to the “greenhouse effect", while increasing levels of particulates (soot, fly ash, volcanic dust) may decrease incoming radiant energy and therefore contribute to decreasing air temperature near the ground. Green peppers were formerly a substantial cash crop in the Annapolis Valley of Nova Scotia, but within the past 10-15 years successive crop failures have resulted in elimination of green peppers as an agricultural product in Nova Scotia. Grapes are, at present, a marginal crop in Nova Scotia. Their sensitivity to climatic change threatens the viability of vineyards in this province. There has been a decrease in the average temperature of the Northern Hemisphere of about 0.3°C, or about 0.5°F since the 1950s (see Mitchell 1975). Mitchell also notes that the effect has been most noticed in the Atlantic sector of the Arctic, and that much of eastern United States has become cooler over the past 15-20 years by more than the global average. 11 TEMPERATURE [ec] ANNUAL MEAN 9.0 ge o + o 6.0 5.0 e MEAN ANNUAL TEMP. @—e PENTADAL AVERAGE bs OO 5 yr RUNNING MEAN 1950 1955 1960 1965 1970 1975 FIGURE 1: Mean annual temperature for Halifax, Nova Scotia (1950-1974) showing pentadal averages and 5-year running means. 12 He notes that summers have become about 1-2°F cooler and winters as much as 4-5°F cooler in some places. At present it is unclear whether the changes noted since the 1940s and 1950s are due to world-wide secular changes in climate or whether they may be due to volcanic dust in the atmosphere (Lamb 1970), atmospheric aerosols (Rasool and Schneider 1971) or changes in the earth's albedo resulting from increased cloud cover as a result of 10-12% higher levels of carbon dioxide in the atmosphere due to fossil fuel combustion. Details of Northern Hemisphere patterns of temperature change are given in Starr and Oort (1973). The principal source of information from which to infer climatic changes are weather records maintained by meteorological stations throughout the world. Very few instrumental records span more than 100 years, and their interpretation is frequently difficult because of increasing urbanization in the vicinity of the instruments, or the relocation of instruments as a result of urban activity. Systematic weather records are available from 1871 in Halifax, although their interpretation is confounded by station relocation in the late 1930s. Nevertheless, interesting trends can be observed from these data. Daily records of maximum and minimum temperatures have provided two measures of climatic trends. Heating degree-days* are computed from the number of days times the number of degrees the temperature is below 18.3°C (65°F). These values are frequently used by fuel oil forecasters in predicting the demand for heating fuel for home or industrial heating. Similarly, the number of degree-days above 5.6°C (42°F) is taken to represent the amount of heat necessary to initiate plant growth*. Based on records maintained at Shearwater Airport in the metropolitan area, the mean annual temperature for Halifax City is shown in Figure 1. Pentadal averages and 5-year moving, averages show a consistent decline of about 1.6°C over the 25-year period since 1950. The decreased mean annual temperature (Figure 1) is converted to total heating degree-days (<18.3°C) (Figure 2). Pentadal and 5-year moving averages are summarized in the trend line, which indicates an increase of approximately 1000 heating degree-days during the period of record at Shearwater Airport. Since 6-8 degree-days is approximately equivalent to one gallon (4.6 litres) of fuel oil for heating purposes, it is apparent that during this * Ed. Note: For more information on heating degree-days and growing degree-days see Hare and Thomas (1974). 13 (ANNUAL) DAYS DEGREE 8500 6500 =! wine jen | = ANNUAL TOTAL DEGREE DAYS HEATING DEGREE - DAYS (<65°F = 18.3° C) HALIFAX, N.S. @—@® PENTADAL AVERAGE Ov 5-yr. MOVING AVERAGE O—A TREND LINE y=-29305 + 18,84x ° ANNUAL AND SMOOTHED fe ee Ce Nr 14 | 1945 FIGURE 2: 1950 1955 1960 1965 1970 Annual total heating degree-days (<18°C) for Halifax, Nova Scotia (1944-1974) showing pentadal averages, 5-year running means and trend line. 1975 period, between 60 and 100 gallons (273 and 455 litres) more fuel oil are required for heating a household to maintain the same temperature as in the 1940s and 1950s. Similarly, Figure 3 shows a decrease of approximately 125 growing degree-days over the same period. Trends over the period of record for the Halifax area are summarized in Figure 4, which treats monthly maximum and minimum temperatures according to a formula developed by Thornthwaite (1948) for calculating potential evaporation. One of the components of the Thornthwaite equation is the computation of the monthly and annual heat index according to the formula: 2 HEAT = z (T;/5) 1-514 (1) where T; = monthly mean temperatures >0°C. Values obtained for the annual heat index are dimensionless, and are normally used in computing monthly and annual potential evaporation from temperature and precipitation data. Alone, however, the annual heat index provides a useful comparative measure comparable to, but independent of degree-day calculations. Figure 4 shows distinct heat index maxima in the 1930s and the early 1950s, anda subsequent decline toward 1970. The cooler years of 1905 to 1930 are shown as a minimum in heat index throughout this period. Also, the last time that the heat index was as low as in the early 1970s was in the early 1880s. These data indicate that recent climatic trends in Halifax are similar in direction and amount to trends recorded elsewhere in the Northern Hemisphere (Mitchell 1975). Increases in heating degree-days and decreases in growing degree-days have serious economic implications. At present, the apparent decline in mean annual temperatures near Halifax since the 1940s implies an increase in home fuel oil consumption of 60-100 gallons (273 - 455 litres) of fuel oil to maintain the same temperature (in the absence of added insulation). Similarly, the loss of green peppers as a cash crop in Nova Scotia reflects the decrease in growing degree-days described in this paper. Although there is no assurance that the present trend will continue, there is equally no assurance that the trend will reverse. It is perhaps appropriate to suggest a workshop or conference of agronomists, foresters, economists, and energy specialists to consider the potential impacts of climatic deterioration or amelioration on agriculture, forestry, 15 1750 (ANNUAL) 1600 =e L Aw y) 2S 1550+ E 1500 | G R E 1450 F D PE 2 ee ee EE nr 1 | eet 1 1945 @—— PENTADAL AVERAGE “of” OO 5-yr. MOVING AVERAGE ; &——— 4 TREND LINE y= 11579-5.05x I ° ANNUAL : fp ti ot er ! | 1950 1955 1960 1965 1970 1975 FIGURE 3: Annual total growing degree-days (>5°C) for Halifax, Nova Scotia (1945-1976) showing pentadal averages, 5-year running means and trend line. 36.00 35.00 + eae 34.00 33.00 32.00 31.00 30.00 28.00 © HEAT INDEX ° 27.00 . @— PENTADAL AVERAGE *. 26.00 Sa a a a | a ae FIGURE 4: Annual heat index for Halifax, Nova Seotia (1874-1972) showing 16 pentadal averages. society, and energy. Contingency plans for crop replacement, altered forestry practices, or energy resource allocation could be developed in anticipation of future climatic trends.* * Ed. Note: It is worth noting that working groups (e.q. on forestry, solar energy, water resources, oceans and fisheries, climatic research, recreation and tourism, transportation, building and construction), co-ordinated by members of the Canadian Climate Centre, have met during 1980. As a result a design document for the Canadian Climate Program will be submitted to the Climate Planning Board in the near future (Canadian Climate Program Newsletter 80-1, January, 1980). Those interested in agriculture are referred to Living with Climatte Change (proceedings of a seminar on food and climatic change held in Winnipeg on January 13, 1977 published by the Science Council of Canada, 1977). Obviously, these seminars and workshops are preliminary to more substantial future contributions in this field. 17 REFERENCES Hare, F.K., and M.K. Thomas. 1974. Climate Canada. Wiley, Toronto. 256 pp. Ladurie, E.L. 1971. Times of feast, times of famine: a history of climate since the year 1000. Doubleday, New York. 426 pp. Lamb, H.-H. 1970. Volcanic dust in the atmosphere; with a chronology of its meteorological significance. Phil. Trans. Roy. Soc. (London), A, 266(1178): 425-533. MacDonald, G.F. 1968. Debert: a Palaeo-Indian site in central Nova Scotia. Nat. Mus. Can., Anthropol. Papers No. 16: 1-207. Mitchell, J.M., Jr. 1963. On the world-wide pattern of secular temperature change. In: Changes in Climate. UNESCO Arid Zone Res. Ser. XX, Paris. pp. 161-181. +. 1975. World climate: major changes under way. NOAA (National Oceanic and Atmospheric Adminstration, Washington, D.C.) Newsletter 5 (2): 1-8. Rasool, S.I., and S.H. Schneider. 1971. Atmospheric carbon dioxide and aerosols: effects of large increases on global climate. Science 173: 138-141. Starr, V.P., and A.H. Oort. 1973. Five-year climatic trend for the Northern Hemisphere. Nature 242: 310-313. Stuckenrath, R.E. 1966. The Debert archeological project, Nova Scotia: radiocarbon dating. Quaternaria 8: 75-80. Terasmae, J. 1975. Notes on climatic change, environment and man. In: Environmental Change in the Maritimes. Edited by: J.G. Ogden, III and M.J. Harvey. Proc. Nova Scotia Inst. Sci. 27 (Suppl. 3). pp. 17-36. Thornthwaite, C.W. 1948. An approach to a rational classification of climate. Geogr. Rev. 38: 55-94. 18 RECENT CHANGES IN TEMPERATURE IN CANADA, AND COMMENTS ON FUTURE CLIMATIC CHANGE M.0. Berry* INTRODUCTION By the early 1980s, it became apparent that a significant cooling of the Northern Hemisphere had occurred over the previous two decades. A number of apparently anomalous climatic events with major negative repercussions were also recorded. One of the most serious of these events, in human terms, was the drought in areas south of the Sahara. These climatic changes or fluctuations provoked considerable speculation regarding their likely persistence and future impacts. Considerable attention was directed to the subject, for example, the Fortune article “Ominous changes in the world's weather” (Alexander 1974). This attention is understandable, since a major change in global temperature, which would almost certainly be accompanied by changes in other climatic parameters such as precipitation, would likely have major repercussions in many countries. For example, in Canada a colder climate could have important effects on key sectors of the economy such as agriculture, tourism, forestry and energy. The objective of this analysis is to examine recent trends in temperature in Canada, and to look at some recent climatic events from a longer-term perspective. TEMPERATURE CHANGE IN RECENT CENTURIES Groveman and Landsberg (1979) have reconstructed the mean temperature of the Northern Hemisphere for the period 1579 to 1976 (Figure 1). For the period 1880 to 1976 the data of Borzenkova et al. (1976) are used. The earlier part of the series is derived from a mixture of instrumental and proxy data, including tree-ring data from Alaska and Finland. While there are a number of uncertainties in the analysis, particularly in the 17th and 18th centuries, it provides a useful index of hemispheric temperature variation. Of * Atmospheric Environment Service, Downsview, Ontario, M3H 5T4 19 1600 1700 1800 1900 FIGURE 1: Reconstructed annual mean temperature departures (from 1881-1975 average) for the Northern Hemtsphere. 1-0 0-25 05 0 0 1900 1920 1940 1960 FIGURE 2: Southern Canadian annual mean temperature trend (1905-1965). 20 particular interest is the indication that the period from approximately 1920 to 1950 was highly anomalous, in that it was considerably warmer than most of the preceding three centuries. After 1950, temperatures appear to decline. This sequence was a basis for concern when the recent cooling trend became apparent, for it appeared that society since the early part of this century may have become accustomed to a climate which was anomalous, and that the anomaly was coming to an end. RECENT TRENDS IN TEMPERATURE A lack of regular weather observations over much of northern Canada prior to the middle of this century has made it difficult to assess, in a comprehensive manner, past variations in Canadian climate. However, Thomas (1975) has prepared a mean annual temperature series for southern Canada, plotted as a succession of 10-year averages, based on data from weather observing stations with reasonably homogeneous records (Figure 2). The series extends back to approximately the beginning of the century and terminates with the 10-year value centred on the mid-1960s. The temperature trend for southern Canada indicated in Figure 2 is roughly similar to the hemispheric trend given in Figure 1, showing a warming early in the century followed by cooling beginning in the 1950s. To what extent has this latter trend been consistent across Canada, and does it appear to be continuing? To partially answer this question Figures 3 and 4 were prepared. These show, respectively, the change in mean temperature across Canada from 1949-58 to 1959-68, and from 1959-68 to 1969-78. Positive values indicate an increase in temperature from the earlier to the later period. The analyses were derived from records of surface synoptic observing stations across Canada. The first period used, 1949-58, corresponds approximately to the decade with the warmest temperature for southern Canada displayed in the Figure 2 time series. The cooling trend appears in the 10-year value centred on 1960. In terms of temperature change from 1949-58 to 1959-68 (Figure 3), the cooling was experienced over much of Canada, however magnitude varied considerably from one region to another. Pronounced cooling (arbitrarily defined as > 0.5°C) occurred in heavily populated southern Ontario, much of Québec, the Atlantic Provinces and part of the Mackenzie-Great Slave region. Over most of the Arctic Islands, a relatively modest degree of cooling 21 CANADA FIGURE 3: Change in mean temperature (°C) from 1949-58 to 1959-68. 22 occurred. In contrast to the rest of the country, much of an area extending from southern British Columbia to southern Saskatchewan experienced pronounced warming (> 0.5°C). Over much of western Canada, the temperature trend reversed in the following period from 1959-68 to 1969-78 (Figure 4). Cooling replaced warming over much of southern British Columbia to southern Saskatchewan. In contrast, a broad swath of warming replaced cooling, from the Mackenzie Delta to northwestern Ontario. Cooling was replaced by little change or some warming in the Lower Great Lakes - Upper Saint Lawrence area, and the Maritimes. Over much of the island of Newfoundland, Labrador, Québec and the eastern Arctic Islands cooling persisted and intensified. In summary, has Canada really become cooler in the 30 years from 1949 to 1978? The eastern half of the country was colder by 1969-78 than in the decade starting in 1949, with the cooling trend in most areas (the Lower Great Lakes, Upper Saint Lawrence and Maritimes being exceptions) intensifying from the second to last decade of the period (Figure 5). In contrast much of the western half of the country was warmer by the third decade, the main exception being an area of cooling in northern British Columbia and part of the Yukon. THE FUTURE While the ability to predict weather or climate for periods in excess of a few days remains modest, progress is being made. On a scale of millennia, prediction has been based primarily on the identification of periodicities observed in paleoclimatic records and extrapolation of these variations into the future. Given the complexity of the earth-atmosphere system and our limited knowledge of the nature and causes of climatic variations, predictions of this nature are, of course, tentative. Berger, et al. (1980), using the last 150,000 years of the paleoclimatic record of Hays et al. (1976), have found a correlation between the series and certain characteristics of the earth's orbit which influence global distribution of insolation, and have periods between 104 and 109 years (Milankovitch Effect*). Extrapolation into SE CRE Pt ee ee ey * Ed. Note: For a succinct discussion of this hypothesis see A. Goudie, 1977. Environmental Change. Clarendon Press, Oxford, pp. 207-210. 23 CANADA le Trout Lake! TN = A exrat# meen à aes : \ on. 15 D Ste A Mew \ wa 5 oe oo iolitan ! ie een Head oy oer oe \: FIGURE 4: Change tn mean temperature (°C) from 1959-68 to 1969-78. 24 CANADA LT Sun RS Sut, Con = the, [Se Bem H) at A Mixer A | 4 ÿ Le se Sa het J RAS oe fn > MP. 20) TT Na ©) < 5 V2) i À | GES re Ot Ve Del FIGURE 5: Change tn mean temperature (°C) from 1949-58 to 1969-78. 25 the future indicates a long-term downward trend in global temperature with successively more severe cooling peaks at 5, 23 and 59 millennia from now. Any trend of this nature would inevitably have many shorter-term variations superimposed on it. One of these would likely be induced by the presently increasing concentration of carbon dioxide in the earth's atmosphere, considered to be caused primarily by fossil-fuel combustion. The best available information suggests that this factor, acting alone, is likely to cause global mean temperatures to increase by 2-3°C in the next 50 to 100 years, with considerably larger amounts of warming at high latitudes. A variation of this nature would be greater than any variation observed in the last 5000 years (Lamb 1980). Other presently unpredictable factors will also influence temperature over the next century, however, assuming the carbon-dioxide associated effect is of the magnitude predicted, the world's temperature should warm significantly. Admittedly, we have a rather modest ability to predict global temperature trends for periods of several decades to millennia. At the other end of the time spectrum, forecasts can be prepared with a significant amount of skill for days, or at best, weeks. For the irvevacdiace period of months to a few decades, prediction is extremely difficult, except in a probablistic manner. So, it is impossible to predict with any reasonable degree of certainty whether the trends in Canadian temperature (Figures 4 and 5) will contime for the next decade or two, or whether they will be replaced by completely different patterns. Over the longer term, by the first half of the next century, an increase in temperature is distinctly possible as a result of increased atmospheric carbon dioxide. However, little is known about the manner in which this warming would vary from region to region, except that magnitude would tend to increase with latitude. Proxy data serving as indices of past climatic variation can play an important role in predicting climate. As more of this information becomes available, and is integrated to provide a coherent picture of historic climate, it will be possible to better define periodicities which may continue in the future, and to obtain a clearer understanding of the causes of climatic variation. In addition, proxy data provide an understanding of the impact of climatic variations on other parts of the environment, and, particularly in areas where there is a lack of instrumental records, aid in defining the range of past climatic variation. Knowledge of such variations in the past can serve as a basis for predicting future climatic fluctuations. Clearly, an ability to predict climate more accurately is needed for design and planning purposes in a variety of human activities. 26 REFERENCES Alexander, T. 1974. Ominous changes in the world's weather. Fortune 89(2): 90-95, 142, 146, 150, 152. Berger, A., J. Guiot, G. Kukla, and P. Pestiaux. 1980. Long-term variations of monthly insolation as related to climatic changes. Presented at the International Alfred Wegener Symposium, Berlin. Borzenkova, I.I., K. Ya. Vinnikov, L.P. Spirina,and P.I. Stechnovskii. 1976. Izmenenie temperatury vozducha severnogo polushariya za period 1881-1975. Meteorologiya i Gidrologiya No. 7: 27-35. Groveman, B.S., and H.E. Landsberg. 1979. Reconstruction of Northern Hemisphere temperatures 1579-1880. Univ. of Maryland Meteorol. Program, Publ. No. 79-181. Hays, J.D., J. Imbrie, and N.J. Shackleton. 1976. Variations in the earth's orbit: pacemaker of the ice ages. Science 194: 1121-32. Lamb, H.H. 1980. The climatic environment of the Arctic Ocean. Presented at the Arctic Committee Conference on the Arctic Ocean: The Hydrographic Environment and the Fate of Pollutants, Roy. Geogr. Soc., London. Thomas, M.K. 1975. Recent climatic fluctuations in Canada. Environment Can., Atmospheric Environment, Climatol. Stud. 28: 1-92. 27 ANALYSIS OF RECENT CLIMATIC CHANGES IN QUEBEC: SOME PRELIMINARY DATA Claude Hillaire-Marcel*, Serge Occhietti**, Louise Marchand* and Reiner Rajewicz* INTRODUCTION A part of a project on the climate of Canada over the past 20,000 years, we were asked by the National Museum of Natural Sciences to carry out a study of recent climatic variations in eastern Canada, essentially on the basis of historical data. The first stage (Hillaire-Marcel et al. 1980) involved planning our research. This paper summarizes results obtained since 1977. Temperature data based on instrumental records for Montréal were compiled for the period 1840-1975 and compared with a corresponding series of records for the Toronto region. In conjunction with this, indirect information on climate, such as port activity at Montréal, was processed and correlated with the instrumental data. Finally, we developed programs for statistical processing and spectrum analysis of short chronological series. These programs allow us to isolate the tendencies underlying the chronological series, and, above all, to bring to light the cyclical nature of recurring events. The programs were tested, as a preliminary step, in the context of the sea level data and recorded levels of inland bodies of water in southeastern Canada. DATA SOURCES Figures 1 and 2 show sources of meteorological information which we have been able to identify, so far, for the Québec City and Montréal regions. Most of the thermometric data for Montréal are from the archives of McGill University. Some gaps (1849, 1853-1862) were bridged by information from the British American Journal of Medical and Phystcal Science and the Canadian Journal respectively. The Québec Meteorological Service has on record * Département des Sciences de la Terre, Université du Québec a Montréal, Montréal, Québec, H3C 3P8 ** Département de Géographie, Université du Québec a Montréal, Montréal, Québec, H3C 3P8 28 “(P) fq syouanol (4) hq pegouep aap s7zva1po1dag ‘puoved ou Jo qavd 18040] smoys quesur ‘reun fi4t9 oeqenb aof vq0p 7v01BOJOU0eZeMW fo SeouN0g :T HUNOIA orsl Oesl oz8l OL8L 0081 06/1 PT TE TT ee te ieee ES OR ee a Re TT hal ket TT | leuunol :(f) VELL ‘AJ & ZELL ‘des ‘8q9nD ap uIseB2W aq - auIze6eW 98qenD aUL (F) —— b : anbipoledics(d) G6LL 118 8 GGZL “Aue! ‘sway Np sinogaq-sauny ayy (fr) — 808L ‘22P & SELL ‘28P ‘(BLE- L6YL) HIDIW AP saalysuy €OSL ‘inf Z ne ‘auel 1a, ‘laps uedaWY using ayy (f) — GO8L 390 LZ ne GOgL ‘Auef 131 ‘Ain912W DeqanD (ft) — 8L8L ‘28P 2 GOSL ‘Aue! “(GLE - L6YL) IDO 2P S2A1491Y QEBL & PZBL ‘sajjansuew sauuadow ZL8L ‘Auel ‘9aqgnH ap ayjazey e7 (r) ‘agen JO A12190S |P9110]SIH pue Aie1a]17 eu JO SUOIJ2ESUEIL (d) 928L sie P GZG8L “AOU ‘28q9nD ap ayjazeyg 27 (tf) — ££BL IAB OE ne ZESL ‘22P HZ (802 D'dV) ‘2qenD ‘sajeuoneu saniysy — GpSL inf e siewu ‘jeushor jeoIskyg pue jedIpay UPOHEUY/ YsNnIg ay) (d) ~ 9G8L IMAP & £G8L ‘22p ‘Eur ueipeueD ayy (d) —_—_————. o_o f—— SZ6L ‘22p eB ZZ8L uni ‘oaqano ‘sajjaanyeu sassaysiy sap 919)SIUIIN “21B0JO108]2W ap 291M8S 29 *(d) liq peqouap eur 87091P01484 -DA0oeu oy JO abd 5017408 SMOYS JUOSUT “Yard 1DZAFUOH aof Dz0p 1v0016oqouoeqeu fo seoanos :T ÆMINOIA 0/81 0981 OS8L ovst SZél ai anbipoisad :(d) v8L 99P 2 ZPBL UNI ‘(2 L- GSZL) IINDIN ap S2AIU91Y gvaL 3des Be pÿ8l UINT'(LZE - LEPL) HIOIW ap saaryuy BPBL “Aa PE GY8L Sie (GL - 68ZL) IIDPW 2P S8A1U91Y LGSL “AOU eB Gÿ8L Sie ‘aouaIDS jedIsAyg PUB jedIpayy JO |BUINOF UPS Ysiig ays (d) 6p8L 12 Sb8L ‘/b8L Sealjansuaw sauuaAow ‘OGgL ‘epeued ‘Ajquiassy aaiejsibay ay} yo jeuinof (d) ZS8L 1008 & 6ÿ8L lew ‘(9 L -682L) IIDIW ep saaryuy ZG8L ‘22P & 6ÿ8L NN! (ZZE- LEVL) IIDIN 8P saaiyry ZG8L 18 LGBL ‘OG8L Sajensuewu sauuadow ‘(p | - 68ZL) II!DN 2P SAIU21Y z98L ‘ides 2 £GgL ‘Auel ‘jeusnor uelpeued au1 (d) G98L IAE @ Z98L ‘1des ‘(Y - 682L) IIDINW 2P saaryqiy 998L seu 8 Z981 ‘Ides ‘(GZE - LPL) I'OON ap serryuy 698L !EW e Gog, leu ‘(g-682L) ON ap S2AIU1Y LZ8L ‘inf eB 8981 ‘AueT"(9ZE - L6tL) ID2W ap saaryouy ee GL6L ‘29P 8 LZ8L ‘inf 2aqenD ‘sajeinjeu sassayoiy sap asajsiuly ‘aibojo1oajayy ap 291A8S ‘(OZE- LGPL) IIDIW ap S8AIU91Y italien fal bt OT TT TT Tt 9Z8L ‘29p 2 ELSL ‘Aue! 30 instrumental data since 1871. Although the data for Québec City are not continuous, they go back to 1792. For the present, only the thermometric series for Montréal has been processed and will serve as a reference (for further details, see Marchand 1979). Port activity (e.g. the dates of the opening and closing of the ports, and of arrivals of the first vessel in spring) was examined in the Port of Montréal archives. Some information concerning the Port of Québec is available from the Archives nationales in Québec City, and some on microfilm from the Public Archives in Ottawa. As the Port of Québec archives were not accessible we refer to those of the Port of Montréal only. Sea level data, and those on inland water levels are available from the Canadian Hydrographic Service, which makes fully updated, computerized archives available to researchers. TEMPERATURES SINCE 1840 Figure 3 shows temperatures recorded since 1840 in Montréal. The coldest year recorded was 1875 and the warmest, 1953. Extreme fluctuations are evident over brief periods - almost 4°C between 1875 and 1877, for example. A comparison of thermometric records for Montréal and Toronto (Figure 4) shows that temperature fluctuations were nearly parallel. Later, we will analyze these data statistically to show differences in range between the two series. The average difference between them, as far as can be determined on the basis of 10-year averages (Figure 5) from 1880 to the present, is about 2°C. Interestingly, this difference is due mainly to low winter temperatures in Montréal. There is not much difference in summer averages for the two cities (Figure 6). Apparently, in these two cities, weather conditions are similar in summer and different in winter. This is probably because heat retained by the Great Lakes in summer is released in winter thus moderating Toronto temperatures then, and because of latitudinal differences between Toronto and Montréal. Further interpretation will be possible once the barometric readings have been compiled and compared to the thermometric series. The period 1860-1874 was an unusual one; temperature fluctuations between Montréal and Toronto show phase fluctuations, or are sometimes intermingled. This is probably the effect of changes in instrumentation (especially in Montréal), in the location of stations, or in SAE OF s1 47 46 44 43 1840 1850 1860 1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 FIGURE 3: Mean annual temperatures of Montréal since 1840. 32 10°C oz ria z = oF ° + Z v Ra Pace ra < => on a aa wae a 5 nena g = eS + ! Cr he ca ~ ees os ap a i ee oe ne Ss Sata awe Se c Sy i = CS Saas i x Crea pen > EE: = Spa FRS. ES SSS, <= —— Sad eg SSS ae Se 2. == ae = a > of Pa 1970 1960 1950 1940 1930 1920 1910 1900 1890 1880 1870 1860 1850 Mean annual temperatures of Montréal and Toronto since 1840. FIGURE 4 33 Lee VAR 44 TORONTO 39 —----- MONTREAL 1840 1850 1860 1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 FIGURE 5: 10-year running means of mean annual temperatures of Montréal and Toronto since 1840. 34 FIGURE 6: 25°C 20 15 10 5 (0) CS i ORONITO MONTREAL -10 Seasonal variations of temperature in Montréal and Toronto based on monthly means (source: Environment Canada). Note that Montréal has colder winters, whereas summer temperatures are similar in both cittes. 35 the recording schedule. For example, six different recording stations can be identified in Montréal during this period. PORT ACTIVITY Figure 7 shows dates marking the opening, closing, and duration of the period when the port was closed, for the Port of Montréal for the years 1870-1915. The choice of a chronological series of this type was dictated by one concern: examination of possible relationships between a series indirectly related to climate and, for example, the instrumentally-recorded temperatures. Although the length of the Port of Montréal's off-season is largely dependent on the climate, it is not a direct reflection of the latter. It depends on: (1) ice thickness in the Saint Lawrence River, which in turn is a direct function of winter temperature; (2) irregularities in the freeze-up and the break-up (essentially downstream from the Port of Montréal), in which a number of variables come into play (e.g. mild, spring-like weather; flooding of tributaries) ; (3) administrative decisions by port authorities. Another indirect indication of the amount of ice in the Saint Lawrence River can be found in the register of the first ship arriving each spring in the Port of Montréal. But, here again, berthing data depend on factors which may be connected as much with the captain's experience as with shipping conditions in the river, or on accidents en route. Above all, technological improvements connected mainly with the advent of the steamship brought about changes in shipping and the possibilities it presented. It should be noted that the climatic significance of this chronological series disappeared as soon as the Saint Lawrence Seaway was opened in 1959, when systematic ice-breaking operations started. There is a poor correlation between the duration of the Port of Montréal's off-season and the temperatures for the corresponding years. However, a comparative analysis of the amount of ice in the river and the winter temperatures must be completed before interpreta- tions are carried any further. 36 suew aiquwaoep ‘AOU eww jiine ‘uel sinol *(1024qu0om ap quod np sea1youy :204N0$) (GI6I-OZ8I) 1024qUon Jo 2404 OY 4D (ydvab doz) uosvas -ffo ayz fo shop ur uozgvanp pur “(ydoab e7ppiu) bursoza ‘(ydpab woqzoq) Buruedo fo se10q :L minota Slél Olél SOél 006! Sé8l 0681 s88l o88l S28 OZ8l 9% Wy : LL ALUN ‘Aou ew praae siew ‘uel aiquiaoep sunol 24nysaAno,p sejeq @unjew4uey ap sejeq ainjawiey ep sveing 37 SEA LEVEL DATA AND INLAND WATER LEVELS As a preliminary measure, records of sea level and water level for 28 stations throughout eastern Canada were compiled (Figure 8). Levels of inland waters (Great Lakes and Saint Lawrence River), which have a climatic connotation more closely related to the concerns of this project, deserve special attention. An example of a tide-gauge record is reproduced in Figure 9. Sea-level, and water-level data characteristically reflect two distinct types of phenomena. The first, without climatic significance involves vertical movements of the substratum, related in particular to the geoidal re-equilibrations (glacio-isostasy, hydro-isostasy, subsidence by sedimentary accumulation, etc.). The second, which is dependent on climate, corresponds, for example, with the effect of storms on sea-level data, and with the amount of precipitation in inland waters. A special mathematical procedure (Rajewicz 1979) makes it possible to isolate the non-climatic phenomena by calculating the tendency of each series (Figure 9). This procedure can be used to gain a knowledge of deformations of the substratum in eastern Canada (Hillaire-Marcel 1979). The residual, or the variations in water level in relation to the tendency identified, can then be interpreted in climatic terms and be subjected to Fourier's series treatment (Figure 10). In this way, cyclical recurrences can be made evident. For example, on the 22-year cycle of high water levels in the Great Lakes (Fairbridge and Hillaire-Marcel 1977), there appear to be superimposed cycles of approximately 6-8 years, 10-12 years and 32-34 years (Figure 11). These cyclical recurrences are observed indistinctly in most series (Figure 12). However, in the Great Lakes, there is a predominance of cycles of 11 years (solar), 18 years (nodal tide), 22 years (Hale's cycle, see Fairbridge and Hillaire-Marcel 1977), and a 33-year cycle, which, so far, is unidentified, but which is probably harmonic with the 11-year solar cycle. Finally, it should be pointed out that some of the major cycles exhibit a phase displacement from one series to another (Figure 13). Fairbridge (1978) has demonstrated that there are relationships between the climatic cycles and longitude, while Schove (1964) found a relationship between latitude and the wave length of the dominant cycles. A more extensive analysis of data gathered in the series we have examined should make it possible to control these geographic variables. 38 FIGURE 8: 500 MILES —+ KILOMETERS WEST LONGITUDE Locations of some of the most complete tide-guage and water-level records in eastern Canada. Key: 1. Churchill, 2. Thunder Bay, 3. Collingwood, 4. Goderich, 5. Port Stanley, 6. Toronto, 7. Kingston, 8. Lanorate, 9. Sorel, 10. Trots-Riviéres, 11. Grondines, 12. Neuville, 13. Québec City, 14. Point-au-Pére, 15. Charlottetown, 16. Halifax, 17. Port aux Basques, 18. Harrington Harbour. 39 FIGURE 9: HALIFAX SECULAR TENDENCY: ~40cm/100 years 1920 1930 1940 1950 1960 1970 1980 Example of a tide-guage series (Halifax), where secondary oscillations of mean sea level are superimposed on a continuous trend -- in thts case a subsidence. 276 FIGURE 10: HALIFAX SCREENED SPECTROGRAM 10. 20. 79 37 Example of a Fourter sertes treatment: record spectrum. : HARMONIC 22 À YEARS the Halifax tide-guage al 42 NIVEAU DE L'EAU DES GRANDS LACS MOYENNES ANNUELLES PRÉCIPITATIONS FIGURE 11: Prectpttatton (top graph) related to water levels of the Great Lakes since 1860. Note the approxtmately 22-year recurrence of high water levels (marked by arrows at bottom). Baste data courtesy of Canadian Hydrographie Service. FIGURE 12: 15 frequency [ Great Lakes Statistical distribution of cycles observed in tide-guage and water-level data for the Great Lakes. 43 = FIGURE 13: Starting years of the most frequent cycles (1900-1980). Note that the eyele seems slightly out of phase from one sertes to the other. Black dots (bottom) represent the 11-year solar cycle. Stars (top and bottom) represent the 45-year Double-Hale cycle. The datum year (1950) ts marked by an arrow. FUTURE PLANS One of the major aims of this type of paleoclimatic research is forecasting future climatic tendencies. There is no use dwelling on the cyclical solar tendencies of 2 and 11 years, which, today, are considered proven and which were confirmed by a preliminary spectrum analysis of the Montréal and Toronto meteorological series. Forecasting of recurrences, especially those of 11 years, can, therefore, be carried out when the phase is determined with precision. Such recurrence is apparently superimposed on a wide sinusoidal fluctuation (Figure 14) the range of which would be about 2°C and the half-wave length, 80 years. This would point toward a moderate cooling off in the climate over the next 50 years, if the temperature maximum observed in the 1950s is used as a basis. It is worth remembering that the existence of such a 160-year cycle was demonstrated by Johnsen et al. (1970) from analysis of ice cores taken during drilling at Camp Century, situated on the western margin of the Greenland Ice Cap. However, we recommend that these forecasts be viewed circumspectly. A parallel relationship between the regular rise in temperatures from the late nineteenth century to the early twentieth century, and an increase in the concentration of carbon dioxide in the atmosphere, connected with industrial activity, can be demonstrated. Nevertheless, we must remember that the increasing "greenhouse" effect, resulting from additional carbon dioxide to the atmosphere, is partly offset by the screening effect of atmospheric pollution. The fact that, since the 1950s, there has been a regular decline in average temperatures also indicates that natural control of the climate goes beyond the human impact. Until there is more information, we can only go by Fairbridge's (R.W. Fairbridge, personal communication, 1979) forecast: "The next century should be cool but not cold". CONCLUSION The above is only a preliminary statement of the project, since temperatures are merely one aspect of climate. Our interpretations will acquire consistency only when we have compared, in particular, the thermometric series and the barometric ones. Moreover, determination of cycles with extended wave lengths requires further work in order to project the series into the past (at least into the eighteenth century). This is the direction we are taking in our work and which we plan to apply to the Atlantic Provinces also. 45 ‘(SL6T-S6/T) (S1odsums fo xequnu qonuun voeu ‘ydoab aeno7) fi1a13on yzodsuns apjos 03 peavduoo (Pesoduraedns aauno 70p1osnu2s payzoows yirn ‘ydvub aeddn) yoeaquoy fo aumanaedueg 1INUUD UDEN :PT MMINOIX 0861 0261 0961 OSél ovél O£6l OZéi Olél 0061 0681 o88l 0Z81 0981 osel Orsi O€8l ozs ols 008! if RE TISSU UF D a BR Es mr \ A | Ost 00z ANNUEL DE TACHES SOLAIRE vin NOMHHI Mi 46 ACKNOWLEDGEMENTS We are grateful to M.K. Thomas and A. Lachapelle (Canadian Climate Centre, Downsview) who provided us with valuable documents concerning, particularly, the Toronto area. We thank R. Lemoine for his analytical support. REFERENCES Fairbridge, R.W. 1978. Climate affected by the core of the earth. Amer. Geophys. Union Ann. Meeting, Miami Beach, Florida. Fairbridge, R.W., and C. Hillaire-Marcel. 1977. An 8,000-yr palaeoclimatic record of "Double-Hale" 45-yr solar cycle. Nature 268:413-416. Hillaire-Marcel, C. 1979. Les mers post-glaciaires du Québec: quelques aspects. D. Sc. thesis, Univ. Pierre et Marie Curie, Paris VI, 2 vols. 600 pp. Hillaire-Marcel, C., S. Occhietti, and G. Prichonnet. 1980. Historical, hydrological and physical evidence of changing climate in Eastern Canada. In: Climatic Change in Canada. Edited by: C.R. Harington. Syllogeus No. 26. National Museums of Canada. pp. 61-72. Johnsen, S.J., W. Dansgaard, H.B. Clausen, and C.C. Langway. 1970. Climatic oscillations 1200-2000 A.D. Nature 227:482. Marchand, L. 1979. Etude comparée des variations de température à Montréal et Toronto depuis 1842. M.Sc. thesis, Univ. du Québec à Montréal. 85 pp. Rajewicz, R. 1979. Analyse des séries chronologiques de l'est du Canada appliquée à l'étude des cycles climatiques. M.Sc. thesis, Univ. du Québec à Montréal. 77 pp. Schove, D.J. 1964. Solar cycles and equatorial climates. Geol. Rundschau 54:448-477. Thomas, M.K. 1975. Dernières fluctuations climatiques au Canada. Environnement Canada, Etude climatologique No. 28:1-92. 47 ANALYSIS OF HISTORICAL EVIDENCE OF CLIMATE CHANGE IN WESTERN AND NORTHERN CANADA A.J.W. Catchpole* and T.F. Ball** INTRODUCTION Analysis of nistorical climatic evidence, now in progress at the Universities of Manitoba and Winnipeg, are based entirely upon Hudson's Bay Company records. The objective is to thoroughly explore the utility of this valuable historical climatic resource before other potential Canadian historical sources are examined (Catchpole 1980). This report deals with three major aspects of this research: (1) a study of ice conditions on northern rivers and seas; (2) a reconstruction of dates of first snowfall and first frost in the Hudson Bay region; (3) the development of a computer-coding system for the retrieval and analysis of climatic information in the Hudson's Bay Company post journals. The immediate goal of this report is to present results which have been obtained to date and to outline the methods of analysis being applied. Interpretation of these and subsequent findings in the light of knowledge of climatic change in the historical period will be presented in a later report. ICE CONDITIONS ON NORTHERN RIVERS AND SEAS The historical evidence of climatic conditions largely comprises descriptive commentaries on weather and environmental conditions contained in written sources. The records of the Hudson's Bay Company are full of climatic information, mainly because they were written in an environment where the vagaries of climate imposed severe restrictions on life. The Company's diarists were not instructed to record weather routinely and systematically, but the rigours of weather in this subarctic environment determined that their diaries frequently mention weather. Casual observers, lacking meteorological instruments, readily perceive those weather elements that are tangible in nature, and references to them commonly occur in their diaries * Department of Geography, University of Manitoba, Winnipeg, Manitoba R3T 2N2 ** Department of Geography, University of Winnipeg, Winnipeg, Manitoba R3B 2E9 48 and journals. Thus, descriptions of ice abound in the historical sources and have proven a rich resource for climatic study. Current research at the University of Manitoba is largely focused on annual fluctuations in ice conditions. The two aspects of ice research to be examined below include: (1) a study of dates of freezing and breaking of river estuaries in the James Bay region; (2) a study of sea-ice intensity in Hudson Strait. The first of these studies is based on information in the post journals, the second utilizes log books of the Company's ships. Dates of Freezing and Breaking of James Bay Estuaries Content analysis has been used to derive dates of first freezing and first breaking of water bodies from Hudson's Bay Company post journals (Moodie and Catchpole 1975). Records of annual dates were derived from information on Churchill Factory (1718 to 1866), York Factory (1714 to 1850), Moose Factory (1736-1870) and Fort Albany (1721 to 1867). Many journals were kept after 1870, but until recently the Company's archival policy prevented consultation of post-1870 journals. Therefore, published records of dates extend from the time of commencement of each journal to no later than 1870. Initial goals of the subsequent phase of the freezing and breaking study were: (1) to derive post-1870 dates from those journals that continue into the twentieth century, thereby linking historical dates with the period covered by modern observations. Fortunately, a change in the Company's archival policy permitted consultation of the post-1870 journals; (2) to extend the network of dated estuaries on the coast of Hudson Bay by deriving dates from Severn House and Eastmain House post journals (Figure 1). When these parts of the study were completed, it was apparent that the trends at Moose Factory, Fort Albany and Eastmain were highly homogeneous. This prompted a third objective - a combination of dates from these three places in order to derive annual means dates of freezing and breaking for the James Bay region. 49 50 100 0 200 300 SOO MILES E 100 © 100 200 300 400 500 600 700 BOO KILOMETRES Churchill Factory an HOUSE \ Norway AUS House ee Fort Alban se MOOtY + i f 4 1 85 60 FIGURE 1: Location map. Post-1870 Dates Of the four journals analyzed previously, Albany alone was sufficiently continuous beyond 1870 to justify an extension of dating into the twentieth century. Albany journals continued to 1921, with only sporadic interruptions. Content analysis was applied without modification to the journals kept between 1871 and 1921. Figure 2 shows dates of first partial freezing, first complete freezing, and first breaking at Albany. With the addition of these dates, the Albany record of freezing and breaking now extends over 200 years. Eastmain and Severn The Eastmain journal spans two periods, 1743 to 1837 and 1893 to 1921. The Severn record commences in 1761 and continues intermittently until 1897. Eastmain House was located on the east side of James Bay and was approximately 230 km east of Albany and 160 km northeast of Moose Factory. It was on the shore of the estuary of the Eastmain River. Severn House was located on the north shore of the estuary of the Severn River, some 22 km inland from the Hudson Bay coast. Content analysis must be specifically applied to each estuary since it is adapted to a consideration of local geographical conditions. Consequently, modified forms of the method were developed for Eastmain and Severn. Modifications involved the identification of five dating places and zones in the Eastmain estuary and four in the Severn estuary. Figures 3 and 4 show the annual dates of freezing and breaking for Eastmain and Severn. James Bay Regional Means There is a high degree of homogeneity between dates derived at Moose and Albany (Moodie and Catchpole 1975). This homogeneity was interpreted as a verification of the validity of these historical dates, but it also indicates that the Moose and Albany dates can be combined to derive means representative of the west coast of James Bay. Similar homogeneity tests were applied to the Eastmain and Severn dates using the Pearson product-moment correlation coefficient. These tests indicate that freezing and breaking dates at Eastmain and Severn are significantly correlated with those at Albany and Moose, but dates of first partial freezing at Eastmain and Severn are not significantly correlated (Table 1). The means, standard deviations and extremes of these dates also demonstrate homogeneity among the three estuaries of James Bay (Table 2). However, the Severn means are substantially different from 51 52 280 Day of the Year 3 O 320 300 320 340 FIGURE 2: 1860 1900 1940 1 1 1 al 1 je 4 First Breaking 120 140 280 First Partial Freezing / | 300 320 First Complete Freezing y 300 i 320 340 1860 1900 1940 Year Dates of first breaking, first partial freezing and first complete freezing at Fort Albany, 1871 to 1940. 120 140 Wee 1 jee 1850 ii 1900 First Breaking m7 First Partial Freezing Wu | 290 | | di | 5 2 | | À | 2 ett | Re 310 \ © > © a 330 T T T T T a ST First Complete Freezing 300 | . 320515 ll 340 1750 1800 1850 Year FIGURE 3: freezing at Eastmain House, 1748 to 1921. Dates of first breaking, first partial freezing and first complete 290 310 330 53 = First Breaking 260 ‘ 4 | First Partial Freezing Day of the Year 280 300 1750 1800 1850 1900 Year FIGURE 4: Dates of first breaking, first partial freezing and first complete freezing at Severn House, 1761 to 1939. 54 TABLE 1: Eastmain Severn Eastmain Severn Eastmain Severn CORRELATION COEFFICIENTS BETWEEN ICE DATES AT SELECTED ESTUARIES. COEFFICIENTS REQUIRED FOR SIGNIFICANCE AT THE 0.01 LEVEL ARE GIVEN IN PARENTHESIS. Moose 0.59 (0.28) 0.47 (0.28) Moose 0.55 (0.30) 0.46 (0.32) Moose 0.69 (0.27) 0.64 (0.28) Albany 0.51 (0.27) 0.44 (0.27) Albany 0.55 (0.27) 0.58 (0.28) Albany 0.69 0.57 (0.27) FIRST PARTIAL FREEZING Severn York 1 York 2 0.36 0.34 0.22 (0.37) (0.39) (0.49) - 0.33 0.45 - (0.49) (0.45) FIRST COMPLETE FREEZING Severn York 1 Yorke? 0.53 0.49 0.46 (0.39) (0.39) (0.57) - 0.61 0.36 - (0.45) (0.49) FIRST BREAKING Severn York 1 York. 2 0.60 0.28 0.56 (0.35) (0.35) (0.45) - 0.57 0.70 - (0.45) (0.42) Churchill 1 Churchill 2 0.12 0.16 (0.56) (0.42) 0.38 Of hr (0.39) (0.56) Churchill 1 Churchill 2 0.40 0.56 (0.37) (0.39) 0.42 0.56 (0.37) (0.48) Churchill 1 Churchill 2 0.34 0.29 (0255) (0.30) 027 0.41 (0.32) (0.32) Note: Definitions of first partial freezing, first complete freezing, first breaking and of York 1, York 2, Churchill 1 and Churchill 2 are given in Moodie and Catchpole (1975). 55 TABLE 2: MEANS, STANDARD DEVIATIONS AND EXTREMES OF ICE DATES. FIRST PARTIAL FIRST COMPLETE FIRST BREAKING FREEZING FREEZING aye. se EE n s e i Dey 8 e i Albany 301 8.0 282 331 313 7.7 291 342 128 5.7 106 150 Moose 304 7,5 281 335 Si. Sion 290 341 127 6.1 105 145 Eastmain 301 6.6 281 328 319 So) 290 348 136 6.1 113 161 Severn 291 7.8 270 311 307 11.2 282 338 132 9.6 ain 154 * n= normal (1747-1776), s = standard deviation (1747-1776), e = earliest whole period of record, 1 = latest whole period of record. Note: 1. Definitions of first partial freezing, etc. are given in Moodie and Catchpole (1975). 2. Dates are given in days after December 31. 56 the James Bay means. These results indicate that it is practicable to combine the Albany, Moose and Eastmain dates in the calculation of James Bay regional means, but that it is not feasible to combine the dates for all estuaries in the calculation of Hudson Bay regional meanSe The James Bay regional means are derived from three records that differ in their terminal dates and continuity. Consequently, annual James Bay means cannot be obtained by averaging three individual values in each year between 1721 and 1921. During that period, for example, are 85 years in which dates of breaking are available from two estuaries, 43 years in which they are available from only one estuary, and eight years in which no dates of breaking are available. In years having individual dates from all three estuaries, the James Bay mean is obtained by averaging these values. When only two dates are available in a particular year, an estimate of the James Bay mean is obtained by adjusting the two available dates and by averaging the adjusted dates. When only one date exists, it is adjusted and treated as the best estimate of the James Bay mean. These adjustments are the amounts by which the 30-year normals of each freezing and breaking date at the individual estuaries differ from the 30-year normals for James Bay derived by averaging the dates from all three estuaries. These normals are based on the period 1747-1776, a 30-year period in which records from the three estuaries are complete. The James Bay regional means are illustrated in Figure 5. Sea Ice in Hudson Strait A previous report (Catchpole 1980) has discussed the potential climatic value of ship's logs preserved in the Hudson's Bay Company archives. These are logs of ships which engaged in the annual fur trade between London and Hudson Bay. Log records extend from 1751 into the twentieth century, being broken only in the years 1839, 1840 and 1841. In most years, two or more logs are available because several logs were often kept on one ship and because the trade was usually conducted by more than one ship. In this initial use of ships' logs, attention is restricted to the westward passage through Hudson Strait. This part of the voyage was selected because of the especially favourable circumstances that it provides for locating the ships precisely. The strait is about 1000 km in length from Cape Digges to 57 Day of the Year 1700 i First Partial Freezing 3104 First Complete Freezing from: nl 3 values : 2values … | j : anne Tvale fi! Ww Means derived 320 1700 FIGURE 5: Mean dates of first breaking, first partial freezing and first complete freezing, James Bay region, 1721 to 1921. 2 58 1 1 1 el ai First Breaking | 1750 1800 1850 Year unning means. Dates are given tn 5-year 1900 Resolution Island and is approximately 200 km wide. The Company's ships followed a route that usually permitted sightings upon landmarks (Figure 6), and these sightings were recorded in the logs so that ships's position can often be located fairly accurately. The Company established a routine whereby trading ships were expected to leave the Thames Estuary on May 31. This was not rigidly adhered to, but ships rarely made their departure more than a week prior to or following the established date. This had the effect, not only of imposing a routine on the voyages (which is beneficial in this research), but also of bringing the ships into Hudson Strait in July when sea-ice conditions were severe and posed a hazard to navigation. An impression of the temporal context of the voyages is given in Figure 7. This diagram shows, in each year, the date of embarkation from London, the period spent in westward navigation of Hudson Strait, the interval in Hudson Bay, the period of eastward navigation through the strait and the time of arrival at London. The logs were written at two-hour intervals with a daily summary. They give routine information on ship's speed, depth of water, wind speed and direction, and weather. A remarks column contains various data including information upon the appearance of ice as well as sailing manoeuvres adopted to avoid ice. A preliminary scrutiny of the logs indicated that they contain two general types of information that may be used to derive indices of annual sea-ice conditions. The first comprises direct descriptions of the ice - its extent, frequency, movement, etc. The second has an indirect bearing on sea-ice severity since it comprises information on sailing manoeuvres required to negotiate sea ice. The logs are full of references to ships being closed-in, or fast, and to occasions when ships were manoeuvred to avoid ice by tacking, grappling or warping. The original objective of this research was to develop a content analysis for deriving indices of ice severity from ice descriptions, but this was later modified to interpret sailing information first. The remainder of this section outlines the method used to derive annual indices of sailing manoeuvres, determined by the presence of ice, during westward passages through Hudson Strait. It is assumed that these indices provide indirect estimates of annual sea-ice intensities. 59 isLANos 2, SALISBURY OF GOD'S °, ISLAND MERCIES NOTTINGHAM ISLAND LAKE HARBOUR ÿ (Lake Inlet?) CHARLES BIG MGS DIGGES ISLAND = ISLANDS... Cape Wolstenholme 283 ISLAND savac Cape Weggs HIGH BLUFF ® OUTER ISLAND ISLAND “WEGGS ISLAND MIDDLE SAVAGE°S\ ISLANDS SADDLEBAC SADDLE ISLAND & LOWER $ ISLAND SAVAGE RESOLUTION = ISLANDS ISLAND Hatton Headland So Rear gy BUTTON ISLANDS ge ÿ a CAPE CHIDLEY QD istanos AKPATOK ISLAND 40 60 80 100 120 140 160Kilometres Se SS FIGURE 6: Hudson Strait with place names mentioned in ships' log books. 60 Dec. 31— 20- 104 Dec. 1 20 : ‘ 1074 . . s . . . Nov. 1 x . a À ° is .<— Return To : es Je oe . hk! AM on 5 Thames Estuary 20 en IF Mes : : as oak 1 sd eae ed he | ; fu | 1 oe i | | ' | | | [es . | 1 Eastward Voyage Oct. 1—.. | = '| 1 | I, ola i| \|' | Through li | ! | | | il '| | | L Hudson Strait 20 É | 10 ty, | ‘| Is | Sept. 1 | | | | | 20 10 ie | / all Westward Voyage | | I |, | | | l' My 1 | Through Aug. 1 Il i | | | | | | | 1] f | | | | ui 1 I Hudson Strait 20 | I 10 July 1 20 Ie : : CRC = Sues tee s+— Leaving . 4 Oo © 25 ANS NC cle tas US oe. +" Thames Estuary 1750 1760 1770 1780 1790 1800 1810 1820 1830 1840 1850 1860 1870 FIGURE 7: Voyages of Hudson's Bay Company shtps between London and the Bay, 1751 to 1870. 61 Data Cards For each year between 1751 and 1870 a data card was prepared and sailing information was recorded on these cards from the logs. Table 3 illustrates the format of the data cards. During westward voyages, the ship was defined as having entered Hudson Strait when Resolution Island was first sighted, and the exit from the strait was indicated by the last sighting of Cape Digges. Daily information was recorded on the cards, for example, names of places visible, distances and bearings of landmarks, references to ice, weather and occasions when people aboard the ship were engaged in trade with the Inuit. Caution was required when recording information on the data cards, since places were sometimes vaguely identified, distances were given in miles or leagues, bearings required correction for magnetic declination and were usually given in the unfamiliar 32-point compass scale (N, N by E, NNE, NE by N, NE, NE by E, ENE, E by N, E, etc.). Durations of Westward Passages Durations varied from more than 50 days to less than 6 days. A wide variety of factors, in addition to the severity of sea ice, are potential causes of such variations in duration. Alternative factors include: weather conditions; ship design and performance; skills and experience of captains; and trade with the Inuit. Furthermore, when the role of sea ice is considered in this context, it is necessary to distinguish between the effects of year-to-year changes in ice severity, and the effects of different times of arrival at Resolution Island. The seasonal sea-ice regime determined that passages which commenced early tended to encounter more ice and therefore were of greater duration. The impact of date of arrival on duration was first studied by correlating the two variables. The Pearson product - moment correlation coefficient was -0.4. This indicates that a weak inverse relationship exists between the two variables. A correlation to actual durations was then obtained using the linear regression relationship: Da = D-b (x;-x) (1) where Da = adjusted duration D = actual duration b = -0.38 xj = date of arrival at Resolution Island in year i x! Il Mean date of arrival at Resolution Island 62 TABLE 3: YEAR: 1772 OFFICER: Je Fowler Day July 30 Aug. 1 10 Place Name Cape Resolution the body of Lower Savage Is. Is. of Saddle Back nearest land S. shore westmost land in sight the body of Great Savage Is. North Shore the body of Charles E end of Salisbury SE end of Nottingham saw Mansfield SHIP: King George Distance NNW1/4W 6 miles ENE 6 or 7 leagues North 10 miles E by N 5 miles NNW1/2W 10 leags. ENE 9 miles East 7 or 8 leags. SWbyW 7 leagues N1/2W 9 leagues NNE 5 leagues SSW to WSW EXAMPLE OF DATA CARD USED FOR RECORDING SAILING INFORMATION. PIECE NUMBER: Remarks haze, cloud fog haze, cloud, rain trade tacked, fog fog tacked for ice, gales, cloud hazy fog, gales rain, gales, heavy showers of snow 376 fog, haze, 63 A second correction for date of commencement of the passage through Hudson Strait involved derivation of the index Dmin. The navigation season was divided into weeks which were numbered sequentially from 1 to 12. Week 1 is June 21-28 (the time of the earliest passage, in 1791) and Week 12 is September 7-13 (the time of the latest passage, in 1811). The total number of passages in each week was enumerated and the minimum duration (Dmin) of the passages in each week was identified. The actual duration (D) in each year was then expressed as a proportion of Dmin for the week in which it occurred. Figure 8 illustrates the temporal variations of actual duration D, and D/Dmin is shown in Figure 9. The relationships between duration and the quality of the ships and abilities of the ships' captains have not yet been fully considered. There are grounds for anticipating that neither of these relationships had an important bearing on variations in duration. An important characteristic of the voyages is that they were often made by several ships in convoy - including, quite often, a Royal Navy escort. Faster ships were routinely delayed while the convoy reassembled. In these circumstances, skillful navigation by individual captains, or rapid sailing of particular ships, were unlikely to have markedly reduced durations. A correlation analysis of the impact of years of experience of captains upon duration has supported this assumption. This was conducted by correlating the mean duration of the voyages undertaken by 24 different captains with the number of voyages undertaken by each captain. This yielded a correlation coefficient of only -0.17. Furthermore, a regression analysis indicated that the durations of voyages under one captain did not decrease as the number of voyages undertaken by that captain increased. Documentation in the logs of weather conditions and trading activities is sufficiently detailed to permit identification of the days on which passage through Hudson Strait was primarily delayed by these factors. Interpretation of this information, and of information pertaining to the effects of sea ice in delaying passages, are presented in the following section. Analysts of Sailing Manoeuvres This analysis commenced with the identification of the days in each year when the following conditions or activities were described in the logs: ship fast in ice; beset by ice; grappling or warping with ice; tacking to avoid ice; adverse weather conditions; and trading with the Inuit. Figures 10 and 11 depict the temporal distributions of occasions 64 Encl d nclosed and Duration Grappling 30 : ; 30 0e Duration : 4 : 25 —— Enclosed and Grappling PRE Ë 25 = 20: 5k 5) 0s on heu se 20 ÿ 2% ete : : ee Zz : tote alle . as : : Ww SA ee 2 2 . oe ee . 2 p =), 15 Fi ie D - a es A hrs 15 o D tas oy “eo? Sper gee Ww : ee an œ ; = 10 10 So 5 O O 1750 1760 1770 /1780° : 1790: 1800 ? 1810 1820 1880 1840/1850" 1860" 1870 YEAR Westward voyages of Hudson's Bay Company shtps through Hudson Strait. The FIGURE 8: annual durations of these voyages are compared with the frequencies with which the ships became enclosed tn ice and grappled with ice. 65 D/Dmin. D/Dmin. FIGURE 9: 66 1750 1800 1850 ] 1750 1760 1770 1780 1790 1810 Year 1800 1820 1830 1840 1850 1860 1870 Annual variations tn the index D/Dmin. The upper graph contains 5-year running means and the lower graph contains actual values. Sept. 20 Day/Month Beset, Completely Stopped /Enclosed by ice, Cannot get Under way.” [4 Fast in ice.” July 1 June 30 Beginning & Ending of Voyages 25 June 19 1750 1760 1770 1780 1790 1800 1810 1820 1830 1840 1850 1860 1870 FIGURE 10: Occasions when shtps were beset by, or fast in, ice during westward voyages through Hudson Strait between 1750 and 1870. The verttcal line shows the duration of the westward voyage through the stratt in each year when the shtps were beset by, or fast tn, tce. 67 Day /Month [] Grappling Beginning & Ending of Voyages 1750 1760 1770 1780 1790 1800 1810 1820 1830 1840 1850 1860 1870 FIGURE 11: Occasions when ships grappled with ice during westward voyages through Hudson Strait between 1750 and 1870. 68 when the ships were beset by and fast in ice, and grappled with ice, respectively. The sailing manoeuvres responsible for delaying progress during prolonged westward passages were studied by correlating actual durations (D) with the following parameters: (1) number of days T on which the ship tacked to avoid ice during the westward passage from Resolution Island to Cape Digges (r = +0.09); (2) number of days G on which grappling or warping with ice occurred (r = +0.84); (3) number of days E on which ship was enclosed in ice, fast or beset (r = +0.50). Both G and E are significantly correlated with duration. The correlation coefficient between D and the sum of G and E (G + E) is +0.92. This indicates that manoeuvres involving contact with the ice significantly prolonged the duration of the westward passage. It implies that the basic cause of annual fluctuations in the duration of westward passages through Hudson Strait was the frequency with which the ship grappled with or became fast in ice. It also implies that annual fluctuations in D provide an indirect indication of yearly changes in the amount of ice. Figure 8 illustrates the relationship between D and (E + G). The occurrence of tacking does not significantly prolong D. Presumably this is because a ship resorted to tacking in the presence of ice which was not of sufficient amount to prevent it making headway. Further, the high correlation between D and (E + G) indicates that factors unrelated to ice play a minor role in determining annual fluctuations in D. Several such factors are under investigation including adverse weather conditions (fog, strong winds, snowfall), as well as delaying activities including trading with the Inuit and waiting for consorts. Preliminary investigations suggest that none of these factors, individually, was a significant determinant of D. DATES OF FIRST SNOWFALL AND FIRST FROST This study has close affinities with that of dates of freezing and breaking. Like the latter it involves the use of information in the Hudson's Bay Company post journals, and its 69 objective is to derive phenological* indices of climatic change. European and Far Eastern research has shown that historical documents containing evidence of climatic change can be most effectively used in two ways: (1) identification of rare and extreme occurrences such as droughts, severe winters, floods or storms; (2) study of phenological indicators such as dates of first snow in fall, river breaking and freezing, first migration of birds or ripening of fruit. The goal of this research is to obtain dates of first snowfall and frost from the post journals. Snowfall and frost studies are being conducted simultaneously using the same general methods. At the outset, these were resolved into the following discrete stages: (1) transcription of journal references to snowfall and frost; (2) development of general methods for analyzing transcribed references, using information from Moose Factory and Fort Albany post journals; (3) testing the results and refining the methods; (4) applying the methods to post journals from a network of trading posts. The procedures which were eventually adopted conformed closely with this scheme, although minor modifications were adopted. The first three stages are now complete. Transcription of References to Snowfall and Frost This work was undertaken by an experienced assistant who had previously transcribed all information used in the study of first freezing and breaking. The objective of the transcription was to extract from the journals, in each year, information that would serve to identify the date of first snowfall and first frost. The assistant was instructed to read each daily entry made between July 1 and October 31, and transcribe verbatim all daily references to snow and frost in that period. For each year, the assistant was permitted to terminate the transcription before October 31 if it became clear to him that the first * Ed. Note: Phenology is the study of relations between climate and recurring natural phenomena. 70 snowfall and first frost had already occurred. In this way, over 200 foolscap pages of daily transcriptions were prepared. Analysis of Transcriptions from Moose Factory and Fort Albany The transcribed information comprised a large amount of heterogeneous descriptions of snow and frost phenomena. Information of this nature had previously been interpreted, in the study of dates of freezing and breaking of water bodies, by the development of a method of content analysis. Previous experience prompted the assumption that a similar method of content analysis would be required to extract dates of first snowfall and first frost from the transcribed information. This endeavour met with less success in the present study than in the previous ones. The following account describes procedures adopted to develop a suitable content analysis of first snow fall and first frost data. Later, utility of the method of content analysis in these contexts will be discussed. One research assistant spent a year seeking to develop methods of content analysis for interpretation of snowfall and frost information. This investigation commenced with extraction, from all transcribed information, of two random samples of daily comments - one pertaining to frost descriptions, the other to snowfall descriptions. Each random sample was scrutinized, and classifications of types of reference to snowfall and frost were developed. The reason for these classifications was to identify types of descriptions of snowfall and frost referring to specific meteorological phenomena and which occur frequently in the post journals. Classes of description which meet these criteria would, it was expected, provide a basis for deriving homogeneous, continuous records of dates of first snowfall and first frost. Initial efforts yielded many classes which were unsatisfactory because they failed to meet the criteria of meteorological specificity and high frequency of occurrence. The classifications were refined to six categories of snow descriptions (snow lying, heavy snowfall, light snowfall, snow with rain, snow with hail, vague references to snow) and four categories of frost description (comments diagnostic of radiation frost, advection frost, recorded sub-freezing temperatures, vague references to frost). AL These classifications of types of reference to first frost and snow are much simpler than the classifications of references to first freezing and first breaking of water bodies developed previously. Several factors have contributed to this contrast. The processes of freezing and breaking of water bodies are protracted, intermittent processes in which several days or weeks elapse between the first appearance of ice or manifestation of breaking, and complete freezing or clearing of ice. The initial objective of the freezing and breaking study was to identify dates of major stages in these processes and the assistant was, accordingly, instructed to transcribe all references bearing upon the various stages. The processes of development of the winter snowpack, and of freezing conditions in the air or in the ground are also prolonged, intermittent processes covering periods of days or weeks. However, at the outset of the snow and frost studies, a limited goal was set of identifying only the initiation of these processes. Therefore, the assistant was instructed to transcribe only data bearing upon their initiation. Hence transcriptions of first snow and first frost descriptions embraced a narrower range of information than the transcription of the first freezing and first breaking descriptions. A second factor contributing to the comparative simplicity of the categories of frost and snow descriptions relates to spatial variations of these phenomena. Within estuaries of rivers draining into Hudson Bay, there are substantial spatial variations in times of river freezing and breaking. These variations are determined by hydrological and topographical conditions, and tend to follow similar patterns from year to year. Therefore, an important component of the classification of freezing and breaking descriptions involved the places to which these descriptions referred. This spatial component has not entered into the classification of first snowfall and frost descriptions, because significant, regular, local variations in these elements are not apparent in the post journals. Testing the Quality of the Dates of First Snowfall and Frost Derived at Moose Factory and Fort Albany The principle of spatial homogeneity testing has been applied in this context. It is founded on the assumption that locations within a particular climatic region tend to experience similar climatic changes. The closer such points are together, and the greater the physical similarity between them, the greater will be the parallelism between climatic changes that they experience. In the study of dates of freezing and breaking this test was applied to dates derived at Moose Factory and Fort Albany. A high degree of homogeneity was 72 observed between the freezing and breaking dates for these posts. This was taken as evidence that the actual dates of these events in the two estuaries are similar and that the dates derived from the post journals are relatively accurate. Similar homogeneity tests were applied in the present study to the following sequence of dates: (1) date of first snowfall where this is defined as the first reference to snow each year, irrespective of the class of information contained in that reference; (2) date of first frost, defined in the same general manner as first snowfall. Annual dates of first snowfall and first frost at Moose Factory and Fort Albany are shown in Figures 12 and 13, respectively. These dates are expressed in days after July 31, and are corrected for both the Gregorian calendar reform and leap years. In the first stage of homogeneity testing, Pearson's product - moment correlation coefficients between dates derived at Moose Factory and Fort Albany were calculated both for the whole period and for 15-year intervals within that period. The results derived from the snowfall dates are shown in Table 4, and indicate a lack of homogeneity between the two estuaries (r = +0.32 in the period 1732-1866). Coefficients derived in the 15-year time intervals demonstrate an erratic variation in the relationships between the two estuaries, with occasions of very close relationships (r = +0.90 in the periods 1792-1801 and 1822-1836) alternating with periods when dates at the two estuaries are unrelated. Figures 12 and 13, and Table 4 show that close parallels exist between the trends at the two estuaries in some periods, but at other times considerable discrepancies exist. The results of this homogeneity test are not decisive. The overall relationship indicates that Moose Factory and Fort Albany are inhomogeneous with regard to dates of first snowfall and frost. However, for brief periods, the two estuaries are much alike. The causes of this inhomogeneity are under investigation and three alternative explanations are being considered: (1) that the post journals are incapable of yielding valid dates of first snowfall and first frost because their authors did not write about these phenomena in sufficient detail; 73 74 FIGURE 12: Fort Albany 201 wtereetsseneees Moose Factory 1705 10 15 20 1725 30 35 40 1745 40 60+ 80 100 20 Days After July 31 40 604. 80 100 i ; 1790 95 1800 05 10 15 20 1825 20 40 60 80 100 Date of first snowfall after July 31, Moose Factory and Fort Albany, 1705 to 1869. FIGURE 13: Days After July 31 Date of first frost after July 31, Moose Factory and Fort Albany, Fort Albany 20 CCETEPEPELEEEEE Moose Factory 40 60 1705 10 15 20 1725 30 35 40 1745 60 80 100 1750 55 60 1765 70 75 80 1785 20 40 60 i 80 100 1790 95 1800 05 10 15 20 1825 20 40 60 80 100 1830 35 40 45 1850 55 60 1865 Year 1705 to 1869. 75 TABLE 4: CORRELATION COEFFICIENTS BETWEEN DATES OF FIRST SNOWFALL AT MOOSE FACTORY AND FORT ALBANY. PERIOD NUMBERS OF CORRELATION PAIRS OF COEFFICIENT DATES 1732-1866 111 +0.32 1732-1746 12 +0.60 1747-1761 15 +0.35 1762-1776 15 -0.40 127779 15 +0.20 1792-1801 10 +0.90 1807-1821 10 -0.02 1822-1836 11 +0.90 1837-1851 12 +0.16 1852-1866 12 -0.25 76 (2) that the post journals are able to yield valid dates, but that discrepancies between tabulated dates at Moose and Albany are caused by errors in transcriptions of references to snow and frost from the primary sources; (3) that post journals are able to yield valid dates and that the discrepancies between the tabulated dates at Moose and Albany have meteorological causes. Accuracy of Transcriptions The accuracy of the original transcribed information was checked by comparing the transcriptions directly with the primary sources. This confirmed the accuracy of the original transcription. A second confirmation of accuracy was obtained by conducting a reliability test in which the dates of first snowfall at York Factory derived in this study were compared directly with similar dates derived by the second author for the same period at York Factory using the same sources (see page 82). Ball's dates were derived from the post journals using the method described later in this report. Identical dates of first snowfall at York Factory were derived by the two studies on 106 out of 139 years. Discrepancies in the remaining 33 years are under investigation: preliminary results indicate that several were caused by simple errors of transcription or coding. The confirmation of the accuracy of the transcription implies that inhomogeneity between dates of first snow and frost derived at Moose Factory and Fort Albany is attributable to the first or third causes listed above. Thus, the possibility arises that the dates derived from the post journals are accurate and that the differences in dates between Moose and Albany have real physical causes. Modern meteorological records are not available from the estuaries of Moose and Albany Rivers. Therefore, a direct investigation of this question cannot be undertaken. An effort has been made to study the matter indirectly by comparing homogeneity between Moose and Albany dates with the homogeneity among similar dates observed at pairs of modern stations located in southern Manitoba. Homogeneity in Modern Observations of First Frost and First Snowfall This study of dates of first frost and first snow observed in southern Manitoba stations cannot confirm the validity of the historical dates derived at Moose and Albany. It is designed to show whether the degree of homogeneity between the historical dates is consistent 1 with the homogeneity between pairs of modern stations near to each other. From the Manitoba network, four stations were selected on the basis of their relative proximity and longevity: Brandon, Minnedosa, Morden and Winnipeg for the study of first snowfall; and Brandon, Morden, Pierson and Winnipeg for the study of the first frost. The homogenity between the dates of first snowfall observed at the four Manitoba stations was tested using the product-moment correlation coefficient. This generated values ranging from +0.16 to +0.46, indicating that the modern data are no more closely related than are the historical dates derived at Moose and Albany. A similar procedure applied to the first frost dates generated correlation coefficients ranging between 0.25 and 0.68. These values considerably exceed the correlation coefficient between historical dates of first frost derived at Moose and Albany. ESTABLISHMENT OF CLIMATOLOGICAL DATA BASE FROM YORK FACTORY AND CHURCHILL FACTORY POST JOURNALS This study is a preliminary phase of the full exploitation of the extensive historical climatological information to be found in the Hudson's Bay Company records. Eventually, it is hoped that all the information contained in these records will be stored in a computer and made available for detailed analysis of the weather patterns of northern North America over the last 250 years. The initial work, briefly outlined here, involves establishment of a system, devised at the University of Winnipeg, for transcribing and coding all references to weather from various types of journals maintained by Hudson's Bay Company employees into a format that allows for easy computer storage and retrieval. Post Journals Consulted, their Durations and Authors The York Factory journal commenced in September 1714 (New Calendar) and continued with interruptions to 1913. The most complete portion of this span occurs between 1714 and 1802, during which only two brief gaps occur. The period from September 1739 to August 1740 is missing because of the loss of the records packet when a ship sank. In August 1782 the Bayside forts were attacked by the French, and a gap extends from that time until September 1783 when the Hudson's Bay Company recommenced business, 78 The post journals were recorded from late August or early September for a complete year. This was the date on which the annual supply ships arrived from England and was the date of any changes in command at the various factories. The meteorological journals, wherein actual instrumental measurements are recorded, also run from approximately September 1 to August 30. System for Transcribing and Coding The objective of the system of transcription was to extract all relevant information in a manner that would facilitate coding of this information for computer storage and analysis. The York Factory journal was transcribed first and in two stages. In the initial stage, all references to weather and weather-related phenomena were listed chronologically and verbatim. This transcription was conducted entirely by the second author. Approximately one year of archival research was required. Following the completion of this stage in the transcription of the York Factory journals, it was assumed that the resultant chronological list of weather descriptions contained a comprehensive sample of types of references to weather and environment that occur in the Hudson's Bay Company post journals. To confirm the assumption that a comprehensive list of weather types had been achieved, a similar transcription was carried out on journals for Cumberland House. These journal were selected because they are from the earliest tnland post, which was established in 1794. A high degree of correlation was found between the terminologies used in the coastal post journals and those used in Cumberland House journals. This is probably because usually men who had maintained the journals at York Factory or Churchill were transferred to Cumberland House. The only differences were found in the use of terms specifically applicable to the proximity of Hudson Bay. For example, references to Sea Roak or Sea Smoke would obviously not be found in the Cumberland House journals. There were no weather types found that were peculiar to Cumberland House. The second stage in the transcription of the York journals involved identification of types of description and the subsequent classification of these using the system contained in Table 5. The Churchill Factory post journal was then coded directly using this classifica- tion without the intermediate transcription of chronological notes that had been applied to the York journals. Coding of the transcribed notes from the York journals, and direct coding of the Churchill information took approximately 2.5 years. 79 TABLE 5: CLASSIFICATION OF WEATHER DESCRIPTIONS IN POST JOURNALS. ELEMENT Precipitation - Rain (27-29)* Precipitation - Snow (27-29) Other Precipitation (27-29) Wind - Direction (23) Wind - Strength (20-21) Other Wind (22) Thermal - Temperature Reading (13-17) Thermal - Relative Warmth (32-33) 80 CLASS Rain(101)**, inclined to rain(102), light(103), drizzle(104), occasional(105), showers(106), squalls(107), heavy(108), continuous(109), most of day(110), part of day(111), with snow(112), at night(113), with fog(114). Snow(201), inclined to snow(202), light (203), small(204), occasional(205), showers(206), squalls(207), heavy(208), continuous(209), most of day(210), part of day(211), with rain(212), with hail(213), at night(214), freezing rain(215), with hail and rain(217). Sleet(301), rime(302), thick rime(303), hail(304), flying showers(305), fog(306), thick fog(307), haze(308), fog in morning(309), mizzling(310), dew(311). N(01), NNE(03), NE(05), ENE(07), E(09), ESE(11), SE(13), SSE(15), S(17), SSW(19), SW(21), WSW(23), W(25), WNW(27), NW(29), NNW(31), Variable (33). Little(01), small(02), gentle(03), light(04), fresh(05), brisk(06), stiff(07), hard(08), heavy(09), strong(10), very strong(11), decreasing (13), moderate(14), storm(15). Calm(1), variable(2), wind(3), breeze(4), gale(5), squalls(6), blowing(7), almost calm(8). Value recorded positive(1), negative(2). Very hot(1), hot(2), very warm(3), warm(4), warmer than yesterday(6), extremely hot(7), cool(8), sharp(9), very sharp(10), cold(11), very cold(12), freezing(13), extremely cold(14), colder than yesterday(15), warm for season(16), cold for season(17), raw(18). TABLE 5: (Cont'd) ELEMENT Thermal - Thaw (35) Thermal - Frost (36) Cloud (30) Thunder (31) Drifting Snow (37) General (34) CLASS Thaw(1), thaw at noon(2), thaw all day(3), thaw in lee(4), thaw in sun(5), thaw out of wind(6). Frost(1), hard frost(2), hoar frost(3), froze hard at night(4), frost at night(5). Clear(1), cloudy(2), overcast(3), flying cloud(4), part clear(5), part cloudy(6), part clear part cloudy(7). Thunder(1), lightning(2), thunder and lightning(3). Drift(1), low drift(2), heavy drift(3). Pleasant(1), fine(2), mild(3), moderate(4), stormy(5), heavy, close, thick, ugly(6), fair(7), sultry(8), variable(9). * Numbers in parenthesis beside each element indicate columns on computer card where element is coded. ** Numbers in parenthesis beside each class indicate the code used to identify that class on computer cards. 81 The coded information was then transferred to computer cards and approximately 6 million bits of information were thus key-punched. Both the transcribed notes and the computer cards were systematically edited in order to identify errors of transcription or key-punching. Analysis of Phenological Indicators Analysis of descriptive climatic information should proceed in two stages. Primary analysis involves the numerical investigation of individual weather elements. Secondary analysis involves the synthesis of many weather elements in the study of synoptic weather systems. Most work to date involves primary analysis. In general, the primary analysis of descriptive climatic information seeks to identify two types of parameters, namely, phenological indicators, and measures of frequencies or intensities of occurrence. Phenological indicators that have been analyzed to date are discussed below. Date of First Rain tn Summer Figure 14 charts data on the first mention of rain after March 1 for each year. Those years in which rain occurred prior to that date are considered as exceptional cases. They will be considered in another portion of the study. Date of First Snowfall in Winter The first snowfall is defined as the first mention of snow after August 1, without distinction between snow flurries or a continuous or heavy fall (Figure 15). Date of First Thaw in Spring The base date for this analysis (Figure 16) was also established as March 1. Records prior to that date will also be shown in a later study. As can be seen from Table 5, there are six classes of thaw with each distinguishing between thawing due to above-freezing ambient air temperatures (thaw in the shade) and thawing due to a positive net radiation balance (thaw in sun). There was no attempt to distinguish between these types in Figure 16. 82 CHURCHILL > œ fe} F ® < i << œ fe) > fter March 1 at York Factory and Churchill (1715-1940). First day of rain a FIGURE 14 83 =" = aks Uv œ 2 TZ U YORK FACTORY AON : 4380190 owron mA Y38W31d3S | isnonv First day of snow after August 1 at York Factory and Churchill (1715-1940). FIGURE 15 84 YORK FACTORY (e) CHURCHILL (a) AVW HOUVW 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 1800 05 10 15 20 25 30 35 1715 First day of thaw after March 1 at York Factory and Churchill (1715-1840). FIGURE 16 85 Date of First Day of Frost in Fall Unfortunately, this is the most incomplete of the graphs (Figure 17) presented, nonetheless trends appear to be evident. The computer was programmed to select the first recorded frost after August 1, but the earliest noted frost is September 1. Interpretation of these data is confused by difficulties encountered in the definition of frost. The problem of identifying frost stems from lack of clear distinction between visible frost and the use of the term frost to merely indicate a relative thermal cooling. In the classification system, frost is placed in two categories, the first under the element of “other precipitation" (Table 5), where it is recorded as rime or thick rime. The second under the element "frost", includes degrees of frost, but also has a category labelled "hoar frost". For the most part, the term rime (also recorded as roak or sea roak) was used to indicate the type of ice that would occur from ice fog drifting in from open water in Hudson Bay. It is important to note that the occurrence of rime is usually associated with an easterly wind. For this reason, rime was classified as a form of advected precipitation, and therefore distinguished from frost occurring due to temperatures falling below the freezing point. A study will be carried out to further distinguish among these various forms of frost and to associate them with other climatic variables. Date of First Night Frost in Fall Unlike the preceding graph, the record of night frost (Figure 18) is restricted to only two categories, namely, "frost hard at night", or "frost at night". Again there is the difficulty of distinguishing hoar frost, or condensed frozen water droplets, from an actual drop in the ambient air temperature below the freezing point. For the time being, it has been assumed that, in both cases, the temperature must fall below the freezing level. However, observers apparently distinguished between precipitated frost and a drop in temperature that was manifested by thin layers of ice on water surfaces in the vicinity. 86 YORK FACTORY (¢) CHURCHILL (4) 35 25 20 15 10 1800 05 95 90 RM HIEW3AON 85 80 75 70 65 60 55 45 40 35 30 25 20 n 5 8 oP 2s IR RAR SNS SEE 1 H380190 | H38W3143S ' ' First day of frost after August 1 at York Factory and Churchill (1715-1840). FIGURE 17 87 YORK FACTORY (e) CHURCHILL (4) 1 1 1 ' 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 1800 05 10 15 1715 First record of night frost after August 1 at York Factory and Churchill (1715-1840). FIGURE 18: 88 Date of First Thunder in Summer A confusing aspect of this phenological indicator (Figure 19) lies in the need for observers to distinguish between thunder and other noises. The sound of ice cracking in the Bay or on the river may occasionally have been mistaken for thunder, particularly if it was a distant sound. The problem is further complicated by the fact that the first thunder would tend to be coincident with the breaking up of the ice, either because of warmer temperatures or because of wind shifts that cause ice to shift. Another possibility is that the sound of ships' cannon, which were used as signalling devices, could be mistaken for a clap of thunder. This is of less concern, because it is possible to check on the arrival or departure of ships around the time of recorded thunder. Finally the assumption was made that lightning and thunder always occur together, even though the observer sometimes only notes one or the other. The term thunder was chosen as the title of Figure 19 because it is the more frequently used term. Presumably, this is attributable to the fact that the ears are "omni-directional", while eyes have a restricted field. The significance of this variable is the association of cumulonimbus clouds with the influx of warm, moist, southerly air. An early migration of the polar front would possibly result in sufficient instability to allow formation of cumulonimbus clouds. It is interesting to speculate on the height of the tropopause* and associated upper air winds during these events, however this will be deferred until the secondary analysis. Date of First Sighting of Geese in Spring This variable (Figure 20) is included as a classic example of phenological information, namely the part played by climate in annually recurring nature phenomena. It is anticipated that changes in the date of first arrival of geese in spring might serve as an indicator of climatic change. The hypothesis derives from studies indicating that the birds migrate when they have a tail wind to assist their movement. As a preliminary test, the graph includes a * Ed. Note: The boundary region in the atmosphere that separates the stratosphere from the troposphere. 89 "(O98I-SIZL) ÂdOFODA YAOX puD 111youny) ‘xeuuns ur depunyz 1541] fo 290q AUO19V4 HHOA —--- THHOYNHOD "60 FANT A isnonv 90 a © = ae 2} su Zz Ô 1715 — CHURCHILL — CHURCHILL YORK FACTORY ---- YORK FACTORY FIGURE 20: Date of first sighting of geese, compared with percentage frequency of south winds, Churchill and York Factory (1715-1826). 91 plot of the percentage frequency of south winds. A cursory study of the graphs suggests that there is homogeneity between the York Factory and Churchill Factory arrival dates. Further, there is an apparent relationship between the graph of first sighting of geese in spring and the occurrence of a higher percentage of south winds. When southeast and southwest wind percentages are also studied, more of the early arrival dates are accounted for. This hypothesis will be tested statistically later. Frequencies of Precipitation Events The frequencies of occurrence of rain in summer and snowfall in winter have been enumerated to date on a monthly and annual basis. Number of Days with Rain in Summer The total number of days on which rain was recorded for the period from June 1 to October 31 is shown in Figure 21. Modern records indicate the Churchill region generally receives rain on a fewer number of days than York Factory. However, it is interesting to note in Figure 21 that the difference between the two stations appears to change through time. Some general observations regarding the trends are noteworthy: (1) The curves apparently follow the same trends with the number of days showing a gradual decrease from 1715 to approximately 1765. From that point the curves rise rapidly to a peak at about 1780, then decline until 1800. Unfortunately, there is a gap between 1800 and 1820, which from all available indications, was an extremely interesting climatic period. When the curves begin again in 1820 they appear to approximate the levels of rainfall previously seen in the period from 1740 to 1750, except that the variability about the mean is greater. (2) The second significant feature of the curves is the separation that begins in 1755 and reaches a maximum difference in 1765. From that year on, the difference remains relatively constant until the available data terminate about 1800. A similar difference is seen between snowfall curves for York Factory and Churchill, however, the separation is not so prolonged, spanning from 1762 to 1780. 92 FIGURE 21: PET oar Numbers of days per year with rainfall, June through October, York Factory and Churchill (1715-1850). 93 1845 The disparity between the two curves in Figure 21 could be attributed to changes in journalists at one or both of the stations. The difficulty with this proposition is that while the journalists changed at both locations none of these changes appears to be coincident with the times of variation in the record. A second possibility is that the changes are attributable to changes in the locations where the journals were kept. York Factory was moved in 1791 a distance of approximately one mile (1.6 km). Churchill Factory occupied two locations, the first being at the original Churchill Factory from 1718 to 1739 and again from 1783 onwards, the second at Fort Prince of Wales from 1740 to 1782. Again, none of these dates are coincident with changes observed in the curves. Number of Days with Snowfall in Winter Figure 22 indicates the total number of days on which snow was recorded for the period from October 1 to May 31. In the computer count of these totals, there was no attempt to distinguish between different types of snowfall. A later count will be carried out to determine the number of days on which heavy or continuous snowfall occurred. Some general observations that can be made on these data are as follows: (1) There is a much greater degree of variation in the number of days of snowfall than there is in the number of days with rainfall. (2) The curves show a gradual decline in the number of days of snowfall over the whole span from 1715 to 1850. This decline is apparent at both locations. (3)Homogeneity is apparent between the two curves, however, it is not as distinctive as that in the rainfall curves. This is probably attributable to the very localized occurrence of snow showers originating from open water in Hudson Bay, particularly in the fall and spring. (4) As previously mentioned there is a separation of the curves between 1763 and 1780. It seems reasonable to suggest that homogeneity of the records is encouraging, and is evidence that the post journals are capable of providing valid phenological and climatic information. Secondly, there is evidence of relatively long-term climatic changes that may be compared to such changes determined by other researchers, both in the region of Hudson Bay and other areas of the world. 94 NUMBERS OF DAYS 80. ~ o 5 > o ty 1715 1720 1725 FIGURE 22: 1730 YORK FACTORY cencee CHURCHILL 1738 1740 1745 1750 1755 1780 1765 1770 1775 Numbers of days per year with snowfall, Factory and Churchill (1714-1852). 1780 1785 October 1790 1795 1800 1820 1825 through May, York 1830 1835 1840 1845 1850 25 Finally, the possibilities of determining meteorological causes for these changes will be greatly enhanced when frequencies of other indicators, particularly wind directions, are plotted and synthesized with the present information. ACKNOWLEDGEMENTS The authors gratefully acknowledge the advice and assistance of Mrs. Shirlee Smith, Hudson's Bay Company Archivist. We also acknowledge receipt of the Hudson's Bay Company's permission to consult and quote from its archives. The first author is particularly indebted to his research assistants Miss Marcia Faurer, Mr. Gerry Madison and Mrs. Mary Magne, each of whom contributed significantly to the work undertaken at the University of Manitoba. REFERENCES Catchpole, A.J.W. 1980. Historical evidence of climatic change in Western and Northern Canada. In: Climatic Change in Canada. Edited by: C.R. Harington. Syllogeus No. 26. National Museums of Canada. pp. 17-60. Moodie, D.W., and A.J.W. Catchpole. 1975. Environmental data from historical documents by content analysis: freeze-up and break-up of estuaries on Hudson Bay 1714-1871. Manitoba Geogr. Studs 5: 1119; 96 REVIEW OF DENDROCLIMATOLOGY IN THE FOREST-TUNDRA ECOTONE OF ALASKA AND CANADA Gordon C. Jacoby and Linda D. Ulan* INTRODUCTION The Tree-Ring Laboratory at Lamont-Doherty Geological Observatory has been engaged in dendroclimatic studies in the eastern United States and in the northern forests of Alaska and Canada since 1976. Tree-ring chronologies developed for these studies are being used primarily for reconstruction of past climatic information. However, archaeologists may also find the chronologies useful. Most of our sampling in the northern forests has been from the forest-tundra ecotone, with the exception of Icy Bay, Alaska in the coastal forest. As part of our efforts, we have reviewed recent activity and past tree-ring studies of the northern boreal forests. For those interested in the coverage to date, compilation of recent collecting and past studies are in Appendices 1 and 2, respectively. Another compilation of such information will be published soon (Cropper and Fritts, in press). Our review is written from a dendrochronological perspective, primarily being interested in old-aged or long-lived trees. We make no attempt to consider the many excellent tree-ring studies relating to forestry and ecology. These studies are often useful to dendrochronologists in trying to understand tree growth and climatic relationships. However, we restrict this review to long-term dendrochronological studies in the forest-tundra ecotone. For background information, we include the following discussion of the general area, past studies, some of the problems and potentials. * Tree-Ring Laboratory, Lamont-Doherty Geological Observatory, Palisades, New York 10964 97 THE FOREST-BORDER REGION For dendroclimatology, the subarctic regions are very important areas in that there are larger, long-term temperature variations in higher latitudes than in tropical or equatorial regions (Lamb 1972, 1977). Temperature is usually the primary climatic variable that influences tree growth in these northern areas. Several of the species that grow in the subarctic region are long-lived and the preservation of dead material is fairly good due to the low temperature in this region. Therefore, the potential is high for development of long tree-ring chronologies for scientific studies in the northern areas. Figure 1 shows the northern tree line and the zones of northern forests. Much tree-ring material has been collected in these areas (Appendices 1, 2). This interest and activity is increasing, and data provided here should not be regarded as all-inclusive. Many foresters, biologists, archaeologists, dendrochronologists and others have collected specimens and/or done tree-ring analyses in the northern areas. Some of these efforts are described elsewhere in this publication by M.L. Parker and his colleagues. Appendices 1 and 2 are merely a first attempt to review work on tree-ring studies of long-lived northern species. For dendroclimatological sampling, the best areas are the limits of species and/or tree distribution, or stress sites within the tree distribution. In the northern forests there are two limits or tree lines: latitudinal and altitudinal. The altitudinal limit ranges from over 2000 m in southern Canada to near sea level in the Mackenzie River delta area. The latitudinal limit ranges from near 55°N at Cape Henrietta Maria on the south shore of Hudson Bay (Larsen 1974) to near 70°N in the Mackenzie River delta area. In places like the Brooks Range in Alaska and the Richardson Mountains in northwestern Canada, the altitudinal and latitudinal tree lines are essentially the same. The present location of the northern tree line approximates the mean summer position of the Arctic front, and there is a theory that this frontal position is what controls the limit of tree growth (Bryson 1966). The configuration of the actual tree line is very complex and defies a simple explanation. Its location involves air masses and their effect on temperature, precipitation, potential evaporation and radiation. Relationships between climatic parameters and the location of tree line are discussed in in Larsen (1974). All large-scale influences are modified by edaphic and microclimatic factors to produce extensions and outliers beyond the regional tree line. These factors are microclimatic zones such as river valleys, sheltered valleys or slopes with a southern exposure, or they can be 98 \75*_65° 170° 70° 165 __ 160° LES ERE COMINANTEN DORE BOREAL FOREST REGIONS OF NORTH AMERICA ee FOREST AND BARREN CANADA a EXT FOREST AND TUNDRA AFTER ROWE 1977 (CANADA) AND VIERECK 1972 (ALASKA) ] ; 150° Es XR FOREST, BARREN, AND TUNDRA-ALASKA 190 0 100,200 300, 400 ee E = EASTERN DECIDUOUS FOREST | MIEES s| id 145° P = PRAIRIE N 0 200 __400 e CHRONOLOGY ; MC=MONTANE AND COASTAL FOREST | KILOMETERS a COLLECTION 135° 130° 125° 120° 115° 110° 105° 100° 95° 90° 652) 80" 7.52 2022. 165° 60° 70° 55° 50° 60° 30° 40° 5 ert “Oa SY SSE EEE EE) Se GE _ aap Gm, GEE A = 50° 5° 55° =< 40° 135° 130° 40° 125° 120° 115° Woe 45° 105° 100° 95° 45° 90° 85° 80° 752 40° 70° 65° FIGURE 1: Boreal Forest regions of northern North America. For details on numbered sampling sites see Appendices 1 and 2. NS, relict stands of trees from an earlier, more favourable period of tree growth (Nichols 1975). From our observations, there are relict stands, areas that appear to be in equilibrium with the northern limit of growth, and areas where new seedlings are establishing themselves beyond a former limit. In addition, because wood is a scarce commodity, in places, people can modify the tree line. Although temperature effects tend to be regarded as the limiting factors for subarctic tree growth and distribution, our results indicate that soil moisture has substantial influence on northern tree growth. Soil drainage varies from very good in sands and gravels of certain areas to extremely poor where impermeable till* blankets the surface, or permafrost underlies it. In sites on well-drained azonal soils or rocky slopes, one would expect moisture stress to become important to tree growth. In poorly-drained areas excessive moisture inhibits tree growth rates, and can even create treeless bogs. Also, in wetter areas heartrot frequently occurs in older trees. Permafrost is a unique feature of the northern portion of the Boreal Forest. Discontinuous permafrost extends through most of that region. In western Canada and Alaska, the forest-tundra ecotone** is sub-parallel to and within the zone of continuous permafrost. Because of permafrost and freeze-thaw cycles in the forest-tundra ecotone, a primary erosional mechanism is solifluction. This can take place on slopes of 15° to 35°, but also even where slope is approximately 2° to 3° (Leopold et al. 1964, p. 346). Solifluctiont often causes "drunken forests", where trees may stand at various angles from vertical. This is an important phenomenon because it is difficult to obtain good dendroclimatological samples in these areas due to reaction wood and abnormal growth in the trees. Micro-site changes due to cryogenic alterations can occur even on stable areas. Small ice lenses can affect the growth of individual trees (Viereck 1965) and complicate ring-width patterns. Ed. Notes: * Boulder-clay deposited by glaciers. ** À transition area between ecological communities. + Downslope flow of soil in a viscous mass. See Brown and Kupsch (1974) for an explanation of terms referring to permafrost. 100 Analysis of tree cores from the Noatak Valley of northwestern Alaska indicates that, during warmer periods, cryogenic effects on tree growth can increase. A location that may otherwise be a good site for tree-ring samples can experience soil movement, tilting of trees and the formation of reaction wood. This abnormal growth presents complications in ring- width analysis. SPECIES USEFUL FOR TREE-RING STUDIES The forest-tundra ecotone contains a limited number of tree species that may be useful for dendrochronology. This limitation is due to low species-diversity and lack of the needed longevity or aborescent form. A life span of at least 200 years is necessary for a species to be of much value. Each species has its own northern limit. A map of the northern limits of Picea, Abtes, Pinus and Larix is provided by Larsen (1974, p. 342). Species distribution maps are available for: Canada (Hosie 1973); Alaska (Viereck and Little 1972); both Alaska and Canada (Fowells 1965). Species which have been used or may be useful in northern regions are listed in Table 1. White spruce (Picea glauca (Moench) Voss) is especially valuable because it grows throughout the northern forests of the Western Hemisphere, and living specimens can exceed 500 years in age - although heartrot is a problem in many older trees (Jacoby and Cook, submitted 1980). Eastern larch or tamarack (Larix laricina (Du Roi) Koch) is also widely distributed and has been used to develop long chronologies (Fritts 1976A). Larch does not seem to reach as great an age as white spruce, but on sites where the two species are mixed larch appears to have more ring-width variation. Long chronologies have been developed from black spruce (P. martana (Mill.) B.S.P.). Its distribution is similar to white spruce, but it is not recorded as being as long-lived. Due partly to the occurrence of black spruce in wetter and less stable areas, it tends to develop reaction wood which interferes with normal growth and clear response to climatic variation. Balsam fir (Abtes balsamea (L.) Mill.) extends through the forest-and-barren area of eastern Canada, but is not reported to achieve great age (Fowells 1965, p. 12). The longer-lived (Fowells 1965, p. 37) subalpine fir (A. lastocarpa (Hook) Nutt.) grows in mountainous areas in western Canada and has been sampled (M. Noel, personal communication, 1978). 101 TABLE 1: * dendroclimatological purposes. LE: AMERICA. COMMON NAME balsam fir* subalpine fir larch, tamarack white spruce black spruce jack pine* paper birch* ** balsam poplar* ** quaking aspen* ** SPECIES Abies balsamea (L.) Mill. Abies lastocarpa (Hook.) Nutt. Larix laricina (Du Roi) K. Koch Picea glauca (Moench) Voss Picea martana (Mill.) B.S.P. Pinus bankstana Lamb. Betula papyrifera Marsh. Populus balsamifera 1. Populus tremulotdes Michx. These species are relatively short-lived: it is uncertain how useful they may be These species are listed primarily because of their wide geographic distribution. dendrochronological potential is low or unknown. 102 POSSIBLE SPECIES FOR DENDROCHRONOLOGY IN THE NORTHERN BOREAL FORESTS OF NORTH for Their Of the North American pines, only jack pine (P. bankstana Lamb.) occurs in large areas of the northern Boreal Forest, but this species seldom achieves sufficient age to be of much use. It is difficult to say if there is much potential for hardwoods. Alnus, Betula, Populus and Salix are widely distributed, but may lack sufficient age to be of much uses DATING OF SPECIMENS FROM THE NORTHERN BOREAL FOREST In general, ring-width variation is less in the northern Boreal Forest than in the south- western United States, where modern dendrochronology developed. Dating is complicated by this relatively low variation. Defining notable rings, or ring patterns, for skeleton-plot type of dating (Stokes and Smiley 1968) may be difficult. For cross-dating, longer sequences may be needed than are required in other areas where there is greater ring-width variation. There is strong serial persistence in most ring-width series from the forest-tundra ecotone. In addition to difficulties of dating trees, dendroclimatologists face problems of a low signal-to-noise ratio for the climatic signal. Archaeologists should be aware that use of tree-ring material for dating in the boreal zones is more difficult than in classic south- western sites. In spite of these difficulties, the potential application in both fields is very good. X-ray densitometry (Schweingruber et al. 1978), which quantifies annual and intra-annual wood-density variations, aids in gaining more climatic information from trees and also greatly improves dating capability. With this new technique, the potential is further enhanced (see Parker et al., this volume). STUDIES BY THE TREE-RING LABORATORY AT LAMONT-DOHERTY GEOLOGICAL OBSERVATORY In 1976 we began collecting tree-ring material in the forest-tundra ecotone of North America. Figure 2 shows the locations of tree-ring sites sampled. As part of a study on earthquake effects recorded by trees, we also sampled a coastal area at Icy Bay, Alaska. The ecotonal collections are almost exclusively white spruce. For the study at Icy Bay, hemlocks (Teuga) and sitka spruce (Picea sttchensis (Bong.) Carr.) were sampled. 103 TREE-RING SITES AND WEATHER STATIONS TT-HH Churchill Arrigetch Creek Sky Pilot Creek Gulf Hazard Dwarf Trees Noatak Sukakpak Mtn. Coppermine Mtn. Sheenjek River Dawson Weather Station Bettles Weather Station FIGURE 2: Locations of tree-ring sites collected by Lamont-Doherty Geological Observatory personnel; and weather stations. 104 A study in Yukon Territory, Canada has resulted in a 400-year chronology that has been used to reconstruct total degree-days above 10°C for June plus July for the Dawson weather station (Jacoby and Cook, submitted in 1980). This reconstruction involved only the ring-width index of that same year, making the chronology and the degree-day reconstruction the same curve with only the scale changed. A ring-width index is a mean value of the standardized annual ring widths for all tree cores for a given oe (Fritts 1976A). The chronology (Twisted Tree-Heartrot Hill, TT-HH) is shown in Figure 3. For this chronology, specimens were collected using increment borers which permit trees to continue growing with only a small temporary injury. The cores were cross-dated using the standard skeleton-plot method (Stokes and Smiley 1968). Ring widths were measured to the nearest hundredth of a millimetre and then dimensionless ring-width indices were obtained using standard tree-ring methods (Fritts 1976A). Trees used in this chronology are all white spruce (Picea glauca (Moench) Voss) from a site along the Dempster Highway, Yukon Territory, in the zone of discontinuous permafrost (Oswald and Senyk 1977). The chronology is a composite of two groups of trees collected along the northeastern and eastern side of an unnamed mountain in this area. One site was called Heartrot Hill due to the prevalence of decayed wood in the interior of most trees cored. We encountered approximately one tree yielding solid wood for every two trees cored. Once we determined a tree was sound enough to produce specimens, we took two to four cores from it. The second location on the more easterly-facing slope was named Twisted Tree for one particularly gnarled specimen that yielded the longest tree-ring series from the site. After dating, measuring, and standardizing ring widths, we found that the two groups of trees were similar enough to be grouped as a single chronology. The final chronology is a mean-value function of 27 core series from 13 different trees. The ring-width patterns, or growth trends were such that most of the standardization was done by a straight line through the mean of the ring-width measurements. For eight cores, a modified negative-exponential curve fit (Fritts et al. 1969) was needed to remove juvenile growth effects. With these curve-fitting techniques, most long-term variations and low-frequency information contained in the ring-width series were preserved in the standardized ring-width indices. Although a few cores went back to approximately 1450, a well-replicated chronology does not begin until around 1550. 105 1550 1600 50 1700 50 1800 50 1900 50 FIGURE 3: Tree-ring chronology for Twisted Tree-Heartrot Hill (TT-HH), Yukon Territory. 106 In order to define the climatic signal in the chronology, response-function analyses (Fritts 1976A) were used. These analyses are an efficient method of comparing tree growth with a large number of meteorological factors in order to identify those which appear most important to growth. The meteorological parameters used were monthly mean temperature, monthly precipitation and degree-days (Petterssen 1969, p. 258) based on records from Dawson, Yukon Territory (Figure 2). For the response-function analysis (Fritts et al. 1971) we used a new program developed in our laboratory. The Dawson weather station is a considerable distance from the site of the chronology, and there are intervening mountains. Therefore, the climatic regions should not be considered as being entirely homogeneous between Dawson and the tree-ring site. Mean annual precipitation could be as much as 25% greater at the site than in Dawson (Oswald and Senyk 1977). Elevation of the Dawson weather station is 324 m and the tree site approximately 900 me Using an average summer lapse rate* for interior Alaska of -0.6°C per 100 m (Haugen et al. 1971), the temperature might be about 3.5°C cooler at the tree site. This rate is close to the standard temperature lapse rate of -0.65°C per 100 m given by Riehl (1972, p. 71). The response-function analysis using monthly temperatures and precipitation is shown in Figure 4. It indicates, as one might expect, that summer temperatures - specifically June and July - were very influential in tree growth at this site. There also appeared to be a dependency on August precipitation. The long-term averages and standard deviation for mean monthly temperatures and precipitation at Dawson are shown in Table 2. It can be seen that the warmest months are June and July. In August, it is still warm enough for some growth processes to occur. Due to low precipitation in this region, apparently above average precipitation is needed in August for cell enlargement and/or cell-wall thickening. In addition to mean monthly temperatures, which are usually used in response-function analyses (Fritts 1976A), we tested degree-days as an alternate predictor of tree growth. The use of heat sums such as degree-days in the study of plant growth and temperature was proposed in 1898 by Merriam (Kramer and Kozlowski 1979, p. 646). Average degree-days for the warm-season months are listed with their standard deviation in Table 2. Degree-day parameters have relatively more variance for warm-season months than mean monthly * Ed. Note: The rate of change of temperature with height in the atmosphere. 107 Response - Function Coefficients 0.1 0.0 -0O.I 0.1 Temperature AMJJAS ON DJ FM AM J J A S Months 0.0 -0.| 108 FIGURE 4: Response-function results for TT-HH using monthly mean temperature and precipitation data from Dawson, Yukon Territory. When the vertical error bars do not reach 0.0, the meteorological parameter ts significant at the 95% level. Precipitation TABLE 2: MONTH Jan. Feb. Mar. Apr. May June July Aug. Sept. Oct. Nov. Dec. = Degree-day data from Haugen et al. (1971) converted to Celsius. ** S.D. = MONTHLY METEOROLOGICAL DATA FROM DAWSON, YUKON TERRITORY TEMPERATURE MEAN -16.8 =25.6 1 we S.D.** standard deviation. PRECIPITATION (mm) TOTAL 19.6 17.0 13.2 11.2 231 33.3 44.7 44.7 32.8 28.0 26.2 25.4 (1921-1965 INCLUSIVE). DEGREE-DAYS* (>10°C) TOTAL 29 30 109 temperatures. The Dawson degree-day data were compiled for 15-day periods from May through August above a base of 10°C and monthly periods above a base of 6°C by Haugen et al. (1971). We analyzed both the 15-day values and monthly values. Using the previously mentioned lapse rate of -0.6°C/100m, the degree-day bases of 10°C and 6°C at Dawson could represent approxi- mately 6.5°C and 2.5°C, respectively, at the tree site. The base for growth degree-days varies for different plant species (Petterssen 1969, p. 258), and tree response to similar heat sums also varies for different species (Hellmers 1962). In addition, the role of "cardinal temperatures", as discussed by Kramer and Kozlowski (1979, pp. 646-647) could interfere with simple temperature-growth relationships. Although it is difficult to estimate the most appropriate degree-day base for white spruce, the best response-function results (highest variance explained) and simple correlations were obtained using monthly values for degree-days above 10°C (Figure 5). In order to use the tree-ring time series to reconstruct meteorological parameters, tree-ring data must be calibrated. Therefore, multiple linear regressions were performed using tree-ring chronology indices for the previous year, the current year, and two following years, to estimate June and July mean temperature and June plus July degree-days. The correlation coefficients (R) and explanation of variance (R2) for the regressions between the tree-ring chronology and the meteorological parameters are listed below: R R2 Ra Degree-days > 10°C 0.49 0.24 0.22 Mean June-July temp. 0.45 0.20 0.18 R32 is adjusted for the loss of degrees of freedom in the regression equations (Neter and Wasserman 1974), and is a more accurate estimate of true variance explained than R2. Although these were multiple linear regressions, the increase in R2 with inclusion of more than the current growth-year indices was not significant at the 95% level. As noted above, with the year of growth as the single independent variable in the regression equation, the "reconstruction" is actually the tree-ring chronology rescaled to equal an estimate of degree-days for June and July. Therefore, Figure 3 can also be looked at as the degree-day reconstruction. 110 Response - Function Coefficients 0.4 0.2 0.0 Degree Days Above 10°C -0.2 M J J A M J J A Months Oz 0.0 Precipitation FIGURE 5: Response-function for the TT-HH dendrochronology site and Dawson City meteorological data, Yukon Territory. 117 In this study, response-function analyses clearly indicate that there is also a significant moisture signal in the chronology. Moisture stress late in the photosynthetic season evidently occurs in other forest-tundra ecotone areas in Alaska and Québec that we have analyzed. From monthly mean temperature and precipitation response-function results (Figure 4), it appears that a colder-than-average April may play a role in tree growth. This pattern has also been observed in our analyses of tree growth in the Brooks Range of Alaska and at the northern tree line in Québec. A cold April may conserve carbohydrates for the coming growing season by maintaining a low rate of respiration. Oxidation of stored carbohydrates (respiration) is strongly influenced by temperature (Lee 1978). Also, a colder April may prevent the trees from breaking dormancy too early and suffering damage from late- season frosts. The moisture signal, and the inverse relationship between tree growth and April temperature, interfere with the pure June-July temperature signal. In spite of these complications, Figure 3 clearly shows effects we attribute to the Little Ice Age (a period of slower tree growth and narrower rings), several warmer decades beginning around 1770 (a period of faster growth and wider rings), a colder period during the mid-1800s (diminished growth rate), and effects of recent Northern Hemisphere warming (increased growth rate). The chronology indicates that this part of the Yukon has recently experienced decades of climate more favourable for tree growth than the long-term mean taken over the last four centuries. We have developed another chronology in the Arrigetch Peaks area of the central Brooks Range, Alaska about 800 km west of the TT-HH site (see Figure 2). The major features of this chronology are similar to the TT-HH chronology as shown in Figure 6. Growth of the Arrigetch trees, however, was not as diminished in the mid-1800s, and response to the recent warming trend lags well behind the response of the TT-HH trees. Analyses of monthly temperature and precipitation data from Bettles, Alaska and the Arrigetch chronology showed the same meteorological factors influencing tree growth as the TT-HH analyses. The TT-HH site lies in a more continental climatic zone than Arrigetch. There is greater seasonal temperature variation at TT-HH and less seasonal lag with respect to warming and cooling at TT-HH (Hartman and Johnson 1978). In 1980, two areas were sampled between these two sites (see Figure 2). Examination in the field indicated strong cross-dating between the Arrigetch chronology and cores from these two new areas. When the new chronologies are developed, we will be able to gain a synoptic view of climate over the northern Alaska-Yukon Territory for the past 400 years. L12 1.4 D Qi Ring-Width Indices arn © w O 00 FIGURE 6: ee ---- Arrigetch 1550 1600 1650 1700 1750 1800 1850 1900 1950 1970 Years Pentads of ring-width indices for TT-HH (Yukon Territory) and Arrigetch (Alaska) dendrochronology sites and changes in mean annual temperature. Temperature data (dotted line) are from Mitchell (1961) for 40° to 70° north latitude using the 1880-1884 pentad as the zero base. (De) LV iis} The northwesternmost trees of North America grow in the Noatak River valley of Alaska. We have sampled three locations in this area (Figure 2), and a preliminary chronology extends to 1515. This ring-width time series is very different from those discussed above. The long-term trends are less well defined. Analysis of these tree-ring series for climatic information awaits the development of a satisfactory final chronology. A large number of samples have been collected from several sites along the Coppermine River in Northwest Territories (Figure 2). The sites are on north-facing slopes, south-facing slopes, fluvial terraces, and eskers. Four sites are in process, and one preliminary chronology has been developed. This preliminary chronology, called Coppermine Mountains, extends back to 1428. It shows large-scale features similar to the TT-HH chronology. Generally, the Coppermine trees seem to be more sensitive climatic indicators than the trees at TT-HH, and statistically the mean sensitivity of this preliminary chronology is higher than at TT-HH. A response-function analysis, using monthly data from the Coppermine weather station, shows climatic influences similar to the TT-HH results. However, there is much more dependence on July temperature alone rather than June and July temperature. Moisture stress is relatively low for this chronology based on data from a bench near the river. A nearby site from the south-facing side of an esker should exhibit a more pronounced moisture signal. Monthly temperature and precipitation data from Coppermine (1933-1978) explain over 34% of the variance in the tree-ring time series. This explanation of variance (R32) is adjusted for degrees of freedom. The unadjusted R2 is 48%. Because the monthly division of meteorological data is arbitrary where growing-season climate is concerned, we feel degree-day information would aid in establishing a closer tree growth/climate relationship in this and other areas. More cores are being processed in order to complete the Coppermine Mountains chronology. Addition of cross-sections from dead trees may extend the chronology to 1340. Our goal for studies in this area is to reconstruct appropriate meteorological parameters for the Coppermine station as far back as we can develop reliable chronologies. We have additional collections from west and east coasts of Hudson Bay. Chronologies have been developed for Churchill and near Gillam in northern Manitoba, and at the mouth of Richardson Gulf in Québec. These chronologies extend back to 1650, 1706 and 1681, respectively. 114 Materials for these chronologies were all collected on brief reconnaissance visits. M.L. Parker of Forintek in Vancouver, and F.A. Hitchcock, formerly of the Museum of Man and Nature in Winnipeg, assisted in some of the sampling near Churchill. Evidently there are older trees near Churchill (Giddings 1954). Our collecting was limited to sites accessible by the few roads at Churchill and near Gillam. Better logistical arrangements and permission to sample trees on the National Research Council site at Twin Lakes would probably result in the production of better chronologies. Sampling at Richardson Gulf was done in cooperation with Jaan Terasmae of Brock University. Several collections have been made by other researchers in this area, and older material has been reported (M.L. Parker, personal communication, 1979). The Hudson Bay chronologies are all influenced by July temperature, and less by June temperature, than the previously mentioned sites in northwestern Canada and Alaska. Weather data from Churchill, Manitoba and Great Whale River, Québec, our three chronologies, and other climatic studies all indicate a later, shorter growing season around the southern half of Hudson Bay as compared to higher latitude regions of western Canada. The region around Hudson Bay is distinct from western Canada and needs further dendrochronological study. We are also collecting other data (e.g. on ice conditions in the Hudson Bay region), and will try to establish relationships between our chronologies and this information. Chronologies from the west side of Hudson Bay may also be of use to archaeologists - particularly those concerned with the history of York Factory. For future studies we have developed an x-ray densitometer system and are beginning to process specimens. With this new method and standard ring-width analyses, we plan to continue dendroclimatic studies along the forest-tundra ecotone, and will try to develop chronologies in a transect across the subarctic area. These chronologies will aid in greater understanding of climatic change along this broad zone where meteorological records are sparse. SUMMARY In the high-latitude forest-tundra ecotone, temperature shifts are larger compared to lower latitudes, and climatically sensitive tree-ring series longer than 500 years can be developed. Long-term climatic data are scarce. This paper reviews recent dendro- climatological studies from the forest-tundra ecotone of Alaska and Canada. 115 In this region, the dating of tree-ring material is more difficult, and the climatic signal less obvious, than in southwestern United States. However, the potential is high for dendroclimatology and archaeology using ring widths and the newer method of x-ray densitometry. Our studies indicate that both temperature and moisture information are recorded in the rings of subarctic trees, especially in summer. Effects of the Little Ice Age and recent Northern Hemisphere warming are shown in past tree growth at several sites. There are also significant differences in sub-arctic tree growth based on data from sites across North America. Efforts are continuing to develop a transect of dendroclimatological sites for the northern forest-tundra ecotone. ACKNOWLEDGEMENTS This material is based on work supported by the National Science Foundation under Grant No. ATM79-19223. Lamont-Doherty Geological Observatory Contribution No. 3089. 116 REFERENCES Blasing, T.J., and H.C. Fritts. 1975. Past climate of Alaska and northwestern Canada as reconstructed from tree rings. In: Climate of the Arctic. Edited by: G. Weller and S. Bowling. Geophys. Inst. Univ. Alaska, Fairbanks. pp. 48-58. Brown, R.J.E., and W.O. Kupsch. 1974. Permafrost terminology. Nat. Res. Council Can. Tech. Memorandum No. 111: 1-62. Bryson, R.A. 1966. Air masses, streamlines, and the Boreal Forest. Geogr. Bull. 8(3): 228-269. Cropper, J.P., and H.C. Fritts. (In press). Ring-width chronologies from the North American Arctic. Arctic and Alpine Res. Drew, L.G. (editor). 1975. Tree-ring chronologies of western America VI: western Canada and Mexico. Laboratory of Tree-Ring Research, Univ. Arizona, Tucson. p. 30. Fowells, H.-A. (compiler). 1965. Silvics of forest trees of the United States. Agriculture Handbook No. 271. U.S. Dept. of Agriculture, Washington, D.C. 762 pp. Fritts, H.C. 1976A. Tree rings and climate. Academic Press, New York, 567 pp. - 1976B. Reconstruction of past climatic variability. Laboratory of Tree-Ring Research, Univ. Arizona, Tucson. p. 64. Fritts, H.C., J.E. Mosimann, and C.P. Bottorff. 1969. A revised computer program for standardizing tree-ring series. Tree-Ring Bull. 29: 15-20. Fritts, H.C., T.J. Blasing, B.P. Hayden, and J.E. Kutzback. 1971. Multivariate techniques for specifying tree-growth and climate relationships and for reconstructing anomalies in paleoclimate. J. Appl. Meteorol. 10(5): 845-864. Giddings, J.L. 1941. Dendrochronology in northern Alaska. Univ. Arizona Bull. 12(4): 1-107. - 1942. Dated sites on the Kobuk River, Alaska. Tree-Ring Bull. 9(1): 2-8. +. 1943. Some climatic aspects of tree growth in Alaska. Tree-Ring Bull. 9(4): 26-32. - 1947. Mackenzie River delta chronology. Tree-Ring Bull. 13: 26-29. + 1948. Chronology of the Kobuk-Kotzebue series. Tree-Ring Bull. 14(4): 26-32. - 1951. The forest edge at Norton Bay, Alaska. Tree-Ring Bull. 18(1): 2-6. +. 1953. Yukon River spruce growth. Tree-Ring Bull. 20(1): 2-5. - 1954. Tree-Ring dating in the American Arctic. Tree-Ring Bull. 20: 23-25. Gray, J., and P. Thompson. 1976. Climatic information from 189/169 ratios of cellulose in tree rings. Nature 262: 481-482. Hartman, C.W., and P.R. Johnson. 1978. Environmental atlas of Alaska. Inst. Water Resources, Univ. Alaska, Fairbanks. 95 pp. Haugen, R.K., M.J. Lynch, and T.C. Roberts. 1971. Summer temperatures in interior Alaska. U.S. Army Cold Regions Res. Eng. Lab., Res. Rept. 244, 39 pp. Hellmers, H. 1962. Temperature effect on optimum tree growth. In: Tree Growth. Edited by: T.T. Kozlowski. Ronald Press, New York. pp. 275-287. aay) Hosie, R.C. 1969. Native trees of Canada. Can. Forestry Serv., Dept. Env., Ottawa. 380 pp. Hustich, I. 1956. Correlation of tree-ring chronologies of Alaska, Labrador and Northern Europe. Acta Geogr. 15(3): 3-26. Jacoby, G.C., and E.R. Cook. (submitted 1980). Past temperature information inferred from a 400-year tree-ring chronology from Yukon Territory, Canada. Arctic and Alpine Res. Kay, P. 1978. Dendroecology in Canada's forest-tundra transition zone. Arctic and Alpine Res. 10(1): 133-138. Kozlowski, T.T. 1971. Growth and development of trees. Vol. 1, Seed germination, ontogeny and shoot growth. Academic Press, New York. 443 pp. Kramer, P.J., and T.T. Kozlowski. 1979. Physiology of woody plants. Academic Press, New York. 811 pp. Lamb, H.-H. 1972. Climate: present, past and future. Vol. 1, Fundamentals and climate now. Methuen, London. 613 pp. - 1977. Climate: present, past and future. Vol. 2, Climatic history and the future. Methuen, London. 835 pp. Larsen, J.A. 1974. Ecology of the northern continental forest border. In: Arctic and Alpine Environments. Edited by: J.D. Ives and R.G. Barry, Methuen and Co. Ltd., London, pp. 341-369. Lee, R. 1978. Forest microclimatology. Columbia Univ. Press, New York. 276 pp. Leopold, L.B., M.G. Wolman, and J.P. Miller. 1964. Fluvial processes in geomorphology. W.H. Freeman and Co., San Francisco. 522 pp. Marr, J.W. 1940. Ecology of the forest-tundra ecotone on the east coast of Hudson Bay. Ecol. Monogr. 18(1): 117-144. Mikola, P. 1962. Temperature and tree growth near the northern timber line. In: Tree Growth. Edited by: T.T. Kozlowski. Ronald Press, New York. pp. 265-274. Mitchell, J.M. 1961. Recent secular changes of global temperature. Ann. New York Acad. Sci. 95:2235-250: Neter, J., and W. Wasserman. 1974. Applied linear statistical models. R.D. Irwin, Inc., Homewood, Illinois. 842 pp. Nichols, H. 1975. Palynological and paleoclimatic study of the Late Quaternary displacements of the Boreal Forest-tundra ecotone in Keewatin and Mackenzie, N.W.T., Canada. Inst. of Arctic and Alpine Res., Occ. Paper No. 15: 1-87. Oswald, E.T., and J.P. Senyk. 1977. Ecoregions of Yukon Territory. Can. Forestry Serv., Victoria, B.C. 115 pp. Oswalt, W. 1950. Spruce borings from the lower Yukon River, Alaska. Tree-Ring Bull. 16(4): 26-32. - 1952. Spruce samples from the Copper River. Tree-Ring Bull. 19: 5-10. - 1958. Tree-ring chronologies in south-central Alaska. Tree-Ring Bull. 22: 16-22. Petterssen, S. 1969. Introduction to meteorology (3rd edition). McGraw-Hill, New York. 338 pp. Riehl, H. 1972. Introduction to the atmosphere. McGraw-Hill, New York. 516 pp. 118 Schove, D.J. 1954. Summer temperatures and tree-rings in North-Scandinavia, A.D. 1461-1950. Geografiska Annaler 35: 40-80. Schweingruber, F.H., H.C. Fritts, O.U. Braker, L.G. Drew, and E. Schar. 1978. The x-ray technique as applied to dendroclimatology. Tree-Ring Bull. 38: 61-91. Stockton, C.W., and H.C. Fritts. 1971. An empirical reconstruction of water levels for Lake Athabasca (1810-1967) by analysis of tree rings. Laboratory of Tree-Ring Research, Univ. Arizona, Tucson. p. 55. Stokes, M.A., and T.L. Smiley. 1968. An introduction to tree-ring dating. Univ. Chicago Press, Chicago. 73 pp. Viereck, L.A. 1965. Relationship of white spruce to lenses of perennially frozen ground, Mount McKinley National Park, Alaska. Arctic 18(4): 262-267. Viereck, L.A., and E.L. Little. 1972. Alaska trees and shrubs. Agriculture Handbook No. 410, Forest Serv., U.S. Dept. Agriculture, Washington, D.C. 265 pp. als) APPENDIX 1: RECENT TREE-RING CHRONOLOGIES OR COLLECTIONS IN THE BOREAL FOREST REGIONS OF NORTH AMERICA. MAP REFERENCE IS FOR FIGURE 1.* | rit SITE NAME SPECIES TIME SPAN LAT.°N LONG. °W REFERENCE : | 1 Barlow Dome Picea glauca 1820-1966 63° 49' 1372832) Drew (1975) | - Chapman Lake à * 1710-1966 64° 51" 138° 791" “ uh 3 Swede Creek " 1800-1966 64° 08' 1s9ce Ast Me D a Dawson Fire Lookout sg 1870-1966 64° 04' 139° 20" " Le 5 Sixty Mile " x 1790-1966 64° 08" 140° 35" ” 1 6 Gold Creek À . 1750-1966 64° 06' 140° 49" m " 7 Bullion Creek m 1690-1966 Gio 02 13837: m ru 8 Lake Beniah " " 1747-1970 63° 29) UPS sea 1710 m ï = Athabasca River Le iW 1708-1970 580022) pS Say Stockton and Fritts (12710 10 Quatre Fourches HY Li 1765-1970 58047 valde. Arey D u " 11 Revillon Coupe * M 1783-1970 58° 52! Tie BF " " " 12 Peace River I de À 1804-1970 538° 58 le ie u LL ve 13 Peace River II 1 1698-1970 Seer 59" dale ABE " T " 14 Claire River à " 1760-1970 58en 530 Udo eM D T Me 15 Thelon Game Sanctuary “ " 1574-1969 63250 1DAN2E Drew (1975) 16 Procrastination Creek " # 1633-1962 Gye aon 142° 30' Blasing and Fritts (1975) 17 Twelve Mile Summit " " 1650-1962 65° 20" 146° 00' wi IL a 120 APPENDIX 1: (cont'd) eee es Se SITE NAME SPECIES TIME SPAN _LAT.°N LONG. °W REFERENCE Salcha River Headwaters Picea glauca 1650-1962 64° 55' 144° 00' Blasing and Fritts (1975) Dawson Junction - ! 1443-1962 64° 10' 141° 30" 2 » = Mount Fairplay se x 1680-1962 63° 50' 142° 00" L n » Hermann's Cabin iy 7 1750-1962 65° 20' 147° 30' = ne Chandalar Lake ys x 1785-1962 6722301 148° 30! " " " Nain Forest (A) a # 1802-1973 56S" 33 62° 00" Fritts (1976B), H.C. Fritts, personal communication, 1978 Nain Forest (B) à Le 1769-1973 56° 33! 62° 00" & ue " Fort Chimo 1-S Picea martana 58° 22' 68° 23' " " " Fort Chimo 2-S us uw 1700-1974 58° 22' 68° 23! i " " Fort Chimo 1-L Larix lartetna 1650-1974 58° 22' 68° 23' m “ " Fort Chimo 2-L by i 1753-1974 58° 22' 68° 23' ue uy " Fort Chimo 3-L “ 1677-1974 58° 22' 68° 23! ud " " Fort Chimo 4-L id " 1641-1974 58° 22' 68° 23! 5 " " Nicol Lake Picea mariana 1873-1974 61° 35' 103° 29' Kay (1978), P. Kay, personal communication, 1978 Slow River M) " 1810-1974 63° 02' 100° 47' LE " 121 APPENDIX 1: (cont'd) MAP REFERENCE SITE NAME SPECIES TIME SPAN LAT.°N LONG. °W REFERENCE 33 Dubawnt Lake Picea mariana 1810-1974 63° 02' 100° 47' Kay (1978), P. Kay, personal communication, 1978 34 Twisted Tree - Heartrot Hill Picea glauca 1459-1975 65° 00' 138° 20' L-DGO** 35 River Crag bs Ue 1635-1975 65° 40' 138° 00" ul 36 Cat Track . hs 1696-1975 65°57" USO" ahs " 37 Arrigetch wh LL 1586-1975 Gyo 7/1 154° 03" A 38 Gulf Hazard ” ¥ 1681-1976 56°) 10 76° 34" " 39 412-Noatak mi Ls 1515-1977 67° 56" 162°) 18% " 40 Coppermine River bd ms 1340-1977 66°-68° 114°-117° uy (several sites) 41 Herring-Alpine Tsuga 1422-1972 60° 26" 147° 45° Fritts (1976B), heterophylla H.C. Fritts, personal communication, 1978 42 Ennadai Lake Picea glauca 1814-1977 61° 101° D. Elliott, personal (several sites) Picea mariana 1793-1977 communication, 1978, 1980! Larix laricina 1638-1977 43 Kasba Lake Picea mariana 1776-1977 60° 06' LOH bis} e " " 44 Picea glauca, 1768-1977 65° 148° G. Juday, personal Picea mariana, 67° 144° communication, 1979 Populus balsamifera 67° 144° 45 Napaktok Bay Picea glauca 1781-1978 SON EGU 62° 35" D. Elliott, personal communication, 1978, 1980! APPENDIX 1: (cont'd) MAP REFERENCE SITE NAME SPECIES TIME SPAN LAT. °N LONG. °W REFERENCE 46 Koroc River Picea martana 1666-1978 58° 04' 65° 02' D. Elliott, personal Larix lariceina 1690-1978 communication, 1978, 1980 47 Churchill Picea glauca 1650-1978 56°45" 94° 04' L-DGO** 48 Sky Pilot Creek " y 1725-1978 56° 24! 94° 22' 4 49 Walker Lake My Ue 1627-1977 67° 00" 154° 00' L. Brubaker, personal (several sites) communication, 1979 50 Border Beacon rn LL 1660-1976 55° 20)" 63215 H.C. Fritts, personal communication, 1978 51 Spruce Creek me us 1570-1977 6s? 31" 138° 40' M. Church, personal communication, 1978 * 52 Mount Sheldon Picea mariana, 1977 62° 130° M. Noel, personal Abies lastocarpa ——— 1977 62° 130° communication, 1978 53 Dwarf Trees (Icy Bay) Tsuga mertensiana 1100-1979 60° 01' 141° 49' L-DGO** Tsuga heterophylla 54 Juneau Tsuga sp. H. Posamentier, personal communication, 1979 55 Narssarssuaq Betula pubescens 1870-1977 61° 09! 45° 22! M. Lawson, personal (Greenland) communication, 1980 56 Quinquadalen : " 1876-1977 60° 16! 44° 30" " mn (Greenland) 57 Edmonton Picea glauca 1882-1969 53° 113° Gray and Thompson (1976) - Okak Bay Area Picea glauca, Se 62° D. Elliott, personal Larix lariceina communication, 1980 123 APPENDIX 1: (cont'd) MAP REFERENCE SITE NAME SPECIES TIME SPAN LAT.°N LONG. °W REFERENCE ** Desulo Laket Picea glauca 56° 03' 63° 48' D. Elliott, personal communication, 1980 Sukakpak Mountain M a 67° 36' 149° 48! L-DGO** Sheenjek River? " " 68° 37' 143° 40' ui Compiled by Linda D. Ulan and Gordon C. Jacoby in 1980, with the cooperation of John Cropper, Laboratory of Tree-Ring Research, University of Arizona, Tucson. Lamont-Doherty Geological Observatory. Not shown on map. 124 | SITE NAME* | White Mountain Alaska Range Noatak (combined) more Creek | 12-Mile Summit Hogatza River re t River Jahl Creek quirrel River ort Yukon-White Eye Stephens Village" alcha Bluff lobe Creek Bluff ower Goldstream riftwood-Series A riftwood-Series B aycock Cleary Summit (combined) = SPECIES Picea APPENDIX 2: TREE-RING CHRONOLOGIES OF NORTH AMERICA (PRE-1960). TIME SPAN 1590-1987 1540-1938 1640-1940 1790-1938 1590-1939 1540-1937 1640-1938 1640-1940 1590-1940 1690-1940 1690-1939 1770-1909 1850-1938 1640-1938 1640-1938 1500-1860 1550-1849 1600-1939 LAT.°N** 66° 64° 67°=682 65° 66° 66° eve ere 67° 67° 6567 66° 64° 65° LONG. °W** 146° 146°-149° 1622-1632 147° 146° 146° 154° 1599 Sine 161° 141°-148° 149° 145° 148° 148° 161° REFERENCE Giddings (1941) 125 APPENDIX 2: (cont'd) ————————…—…—…—…—…"…"…"…"…"—…"—"…"—"—" "—"—"—"—"—"—"—"—"—"— SITE NAME* SPECIES TIME SPAN LAT.°N** LONG.°W** REFERENCE L Koyuk (Norton Bay) (combined) Picea 1640-1940 65° 161° Giddings (1941, 1953) | Great Whale River Picea glauca, 1765-1939 55° 77° Marr (1940) Picea mariana Gulf Hazard Picea glauca 1700-1939 56° 76° T 2 Moose River Picea martana 1800-1946 52° 81° Hustich (1956) Great Whale River Mt " 1800-1946 559 Ti L " Knob Lake ci . 1800-1946 55° 67° Ls " Knob Lake Picea glauca 1800-1946 55° 67° a" at Mecatina Picea mariana 1817-1946 53° 63° T iy Nulato Picea 1597-1947 65° 158° Oswalt (1950) Kaltag " 1710-1947 64° 1592 " " Anvik 1733-1947 63° 160° " " Holy Cross ag 1622-1947 62° 160° " " Russian Mission x 1737-1947 62° 161 Ve Marshall We 1703-1947 65° 162° " LL Shaktoolik " 1700-1949 65° 160° " LL Paxson " 1750-1940 63° 146° Oswalt (1952) "Menasta Group" ou 1755-1950 63° 144° " Mi Slana Group is 1700-1950 63° 144° " " 126 APPENDIX 2: (cont'd) SITE NAME* Gakona Chitina Moody Cantwell Chulitna Talkeetna Eureka Atlasta Beaver Groups Purgatory Group Stevens Village Group Rampart North Group Tanana Group Birches Group Kokrines Group McGrath, Kuskokwim Group Nulato Group Circle Group SPECIES Picea Picea glauca LL LL LL LL LL Picea w" " LL LL LL " LL LL " TIME SPAN 1780-1950 1723-1950 1840-1952 1762-1952 1861-1952 1854-1952 17911952 1781-1952 1802-1941 1740-1941 1675-1941 1730-1941 1730-1941 1700-1941 1720-1941 1700-1941 1700-1941 1740-1941 LAT. °N** 62° LONG. °W** 145° 145° 149° 149° 149° 150° 147° 146° 147° 148° 149° 150° 1522 153° 155° 156° Het hed 144° REFERENCE Oswalt (1952) Oswalt (1958) LL LL LL " Giddings (1943) LL LL Giddings (1943, 1953) 127 APPENDIX 2: (cont'd) SITE NAME* SPECIES TIME SPAN LAT.°N** LONG.°W** REFERENCE Mackenzie River Delta Picea 1357-1945 67°-68° 134°-135° Giddings (1947) Kobuk-Kotzebue Seriest D 978-1947 Giddings (1948) Cape Darby . 1527-1949 65° 163° Giddings (1951) Kobuk River-Amber Island # ne 1700-1760 Giddings (1942) * These are named groups of trees used to form time series of ring-width data. The groups are often different from what is termed a site in current literature. Ring-width data may not extend to the age of the oldest trees. Other collections are mentioned by Giddings, but no associated ring-width information has been published. Beginnings of time spans are estimates by Giddings. ** Latitudes and longitudes are estimated from maps in the various references and only serve as a guide to general locations. + Includes archaeological material and driftwood. # Other archaeological material was collected in this area, but has not been dated absolutely. 128 DENDROCHRONOLOGICAL STUDIES ON THE COASTS OF JAMES BAY AND HUDSON BAY M.L. Parker, L.A. Jozsa, Sandra G. Johnson and Paul A. Bramhall* INTRODUCTION Dendrochronological samples were collected from living trees along the east coasts of James Bay and Hudson Bay in June, 1979, to build tree-ring width and density chronologies for proxy climatic data of that region. This field collecting was sponsored mainly by the Paleobiology Division, National Museum of Natural Sciences, National Museums of Canada. C.R. Harington, of the Paleobiology Division, is coordinating a project to reconstruct climatic history in the Hudson Bay region. His approach is to use data from various sources, such as geological, hydrological, dendrochronological and historical records. One specific and immediate objective is to compare proxy climatic data obtained from historic Hudson's Bay Company records with proxy data derived from tree-ring series. Tree-ring samples also were collected later in the month of June on the west coast of Hudson Bay, with Gordon C. Jacoby of the Lamont-Doherty Geological Observatory of Columbia University, and Ann Hitchcock, then of the Manitoba Museum of Man and Nature. Additional information about these collections is presented by Jacoby in this publication. Developments in the field of dendrochronology during the past two decades, particularly the technique of x-ray densitometry, have greatly improved the potential for using tree rings for reconstructing climatic history (Parker and Kennedy 1973; Parker 1976). The tree-ring samples obtained for this project are being processed at the Western Forest Products Laboratory, Forintek Canada Corporation, using the technique of x-ray densitometry. * Forintek Canada Corporation, Western Forests Products Laboratory, 6620 N.W. Marine Drive, Vancouver, British Columbia, V6T 1X2 129 The results of this study will be presented in parts, as the processing of the material from the various tree-ring sites is completed. Two parts are presented here: Part 1 describes sites for tree rings collected on the International Union for Quaternary Research (INQUA) Hudson Bay Field Meeting, and Part 2 presents the tree-ring data from one of those sites, Cri Lake*, in the Great Whale River (Poste-de-la-Baleine), Québec, area. Subsequent reports will present the tree-ring data and additional site information for the other sites described here and from the sites located on the west side of Hudson Bay. Tree-ring samples from the sites described in Parts 1 and 2 of this report were collected by the first and second authors. We were fortunate to have been able to accompany a group of geologists on the INQUA excursion to areas where we wished to obtain tree-ring samples. Additional information on climatic history will be available when tree-ring data have been derived from all of these sites. This approach can be viewed as a long-term plan to develop a network of tree-ring sites throughout the northern portion of the Boreal Forest - a region of significance for climatic studies. * Ed. Note: This place name is not listed in the latest edition (1978) of the Gazette officielle du Québec, and presumably is only used locally. 130 PART 1: FIELD COLLECTING FROM MATAGAMI TO RICHMOND GULF, QUEBEC L.A. Josza and M.L. Parker BACKGROUND Tree-ring samples were collected from four major sites in Québec along the coasts of James Bay and Hudson Bay in June, 1979. The field collecting trip was made in the company of geologists associated with the INQUA Hudson Bay Field Meeting. The first segment of the trip was the James Bay excursion from Matagami to Radisson. The second part of the trip was to the east coast of Hudson Bay, and centred on the town of Great Whale River, Québec. Our purpose in joining this field excursion was to draw on the knowledge of regional experts in geology and general forestry concerning coastal areas of James Bay and Hudson Bay in order to collect tree-ring samples covering approximately the past 300 years. The tree-ring proxy data are required for checking and verifying historical records that were kept by the Hudson's Bay Company posts in this region. GENERAL DESCRIPTION OF THE REGION The region's forest cover and vegetation from Matagami, in the "closed" Boreal Forest zone, to Richmond Gulf, near the northern treeline, is both topographically and climatically controlled. Due to the presence of Hudson Bay and the resulting cooling effect during the growing season, tundra vegetation is found here farther south than anywhere else in Canada. The book Forest Regtons of Canada (Rowe 1972) provides both a general and a detailed description of the forest geography and plant communities throughout Canada. Each region in Rowe's classification is considered as a major forest type, characterized by a broad uniformity in tree species composition without ecological implications. Of the eight regions described by Rowe, the Boreal Forest Region comprises the greatest part of the forested area of Canada, forming a continuous belt from the East Coast westward to the Rocky Mountains and northwestward to Alaska (Figure 1). 131 SFR EN Sr ~ Boreal Forest Region of Canada (Rowe 1972). FIGURE 1: 132 Rowe's classification distinguishes three zones within the Boreal Forest Region: (1) Predominantly forest (the Matagami Airport tree-ring site in Québec is in this zone); (2) Forest and grassland (transition to the prairie in Alberta, Saskatchewan and Manitoba); (3) Forest and barren (the Radisson, Cri Lake and Richmond Gulf tree-ring sites in Québec are in this zone). Rowe further describes a more detailed subdivision of the regions, resulting in 90 forestland units across Canada called "Forest Sections". Of these, only three have relevance to the Québec sites where tree-ring samples were collected during the INQUA excursion (Figure 2): (a) Forest-Tundra (Cri Lake and Richmond Gulf); (b) Fort George (Radisson); (c) Northern Clay (Matagami). All of the 90 sections are conceived and described as geographical areas having their own individuality, with distinct patterns of vegetation and physiography. (a) The Forest-Tundra has a broad geographic range, but the climatic severity and the presence of continuous permafrost has produced a similarity in vegetational structure throughout. The pattern shows a gradual change from south to north as forests shrink and tundra expands. The principal tree species are black spruce (Picea martana (Mill.) B.S.P.), white spruce (Picea glauca (Moench) Voss) and tamarack (Larix laricina (Du Roi) K. Koch), accompanied by species of alder (Alnus), birch (Betula) and willow (Salix) shrubs. White spruce, in this section, is found only in a narrow coastal strip around the southern margin of Hudson Bay and along the Atlantic Coast, due to the maritime influence. Farther inland, black spruce and tamarack dominate the wooded patches. Black spruce trees frequently assume "candelabra" form as the treeline is approached. The widespread presence of permafrost (Figure 3) and the low soil temperatures result in a shallow root zone. Exposure to wind, low air temperatures during the short growing season, and soil instability all combine to limit tree growth. (b) The Fort George Forest Section is located east of James Bay on a lowland, covered by glacial till or boulder clay, with a gradual upward slope to the east. The topography varies 3s FIGURE 2: AA ee Z Vij a / Forest Sections relevant to the Québec tree-ring sites mentioned in the text: a. Forest-Tundra, b. Fort George, e. Northern Clay. SCALE 200 0 200 400 600 800 100 - Kilometres ! ! ! 1 ! ! ! ! | Continuous permafrost Widespread permafrost Southern fringe of permafrost region FIGURE 3: Permafrost Regions of Canada (Rowe 1972). 135 from flat to rolling and, although there are some rugged places, relief is generally low. The land is a patchwork of bedrock outcrops, peat bogs, sand and gravel. Black spruce is the dominant tree species on shallow organic soils, occurring in both open and dense stands. Strips of tamarack frame the lakes and water courses. Jack pine (Pinus bankstana Lamb.) is abundant on higher ground, and shows vigorous growth on sandy and gravelly soils. Jack pine stands are even-aged, and the younger stands show evidence of fire. This section is situated in the southern fringe of the permafrost region. (c) The Northern Clay Forest Section has a nearly level topography with surface deposits of water-worked glacial tills and lacustrine materials from Glacial Lake Ojibway. Extensive stands of black spruce dominate the landscape. Tamarack is found on lowland wet sites in association with young stands of black spruce. Improved drainage is reflected in hardwood or mixedwood stands of trembling aspen (Populus tremulotdes Michx.), balsam poplar (Populus balsamifera L.), balsam fir (Abtes balsamea (L.) Mill.), white spruce and black spruce. Jack pine is the principal tree species on the drier sites, such as outwash deposits, old beaches and eskers*. White birch (Betula papyrifera Marsh.) also is prominent on sandy soils. The absence of surface rock, the extensive poorly-drained flats, relatively few lakes, and clay-banked gently-flowing streams are the main characteristics of this section. This whole section is well south of the southern fringe of the permafrost region. TREE-RING SAMPLING SITES Matagami Airport Site, Québec The Matagami Airport tree-ring site, in the Northern Clay Section, is located 13.8 km southwest of Matagami, Québec, beside the road to Matagami Airport (Figure 4). The principal tree species encountered is black spruce growing in cool, wet sites with lichen and moss ground cover (Figure 5). The southernmost tree-ring collection was on a site near Matagami * Ed. Note: Long narrow ridges of sand and gravel deposited by streams flowing on, in or beneath the ice of a glacier. 136 FIGURE 4: HUDSON Great Whale River SCALE 100 0 100 eu | KILOMETRES Radisson b Matagami Airport Site ee 3@Val-d Or INQUA excursion. Sites in Québec where tree-ring samples were collected on the 1979 ts quite dense. Note that the stand tte. Matagami tree-ring st FIGURE 5 138 Airport, which can be described as a "closed" Boreal Forest cover zone with 95% black spruce. Twelve black spruce trees were sampled, 11 with a Swedish increment borer, while one tree was felled and a complete cross-sectional disk was obtained. The collection date was June 16, 1979. On the average, trees were small and slow-growing on this site, with a mean diameter at breast height (1.3 m above ground) of 18.5 cm, ranging from 13.7 to 23.1 cm, a mean height of 18 m, and 18 annual rings per cm. The oldest tree was 195 years old. Growth trend was generally flat or reversed. This resulted in a uniform ring-width series from pith to bark, or the tree produced narrow rings initially (near the pith) and then progressively wider annual rings to an average level that was then maintained for the life of the tree (Figure 6). Areas of compression wood indicate several instances of leaning of the stems, possibly caused by wind action and frost heave. Radisson Site, Québec The Radisson tree-ring site (Figure 4) is located approximately 90 km from Hudson Bay and 5 km south of La Grande Riviére, also known as Fort George River (Figure 7). This area is situated in the northern part of the Boreal Forest, characterized by open stands of small conifers, mainly black spruce (Rowe 1972). An almost pure stand of black spruce, with sporadic jack pine, was sampled on a north-facing slope on June 20, 1979. The grade was approximately 50%, relatively steep, with large exposed rock outcrops. The hillside also was strewn with large boulders. Ground cover was mainly moss, lichen and Labrador tea. Some of the trees were scarred and an attempt will be made to date the event that caused this. Identical calendar dates of tree rings of several trees could indicate fire scarring. Fifteen black spruce trees were sampled; 11 with a Swedish increment borer and four trees were felled to obtain complete cross-sectional disks. The average diameter at breast height for the sample trees was 16.5 cm, ranging from 12.4 to 23.4 cm. The stand was only 105 years old and approximately 8 m tall. The ring-width pattern from pith to bark was mainly uniform, with a few examples of suppressed growth in the pith area. The tree-ring samples were free of compression wood, indicating stable ground conditions. 139 FIGURE 6: Commercial timbers cut from black spruce trees in the Matagamnt area, Québec. Note that the rings are narrow and that they vary little in width from ptth to bark. 140 FIGURE 7: Radisson tree-ring site in foreground and topography of the area looking north (distance). 141 Ÿ i a . marian > % CN & = { ; Kae nm section of Sample #9, from Radisson, Québec, and the top of the a o Cros FIGURE 8: croun of that tree showing the stage of flushing on June 20, 1979. 142 The site appeared to be well drained and there was no evidence of frost heaving that typified most black spruce stands on wet sites in this area. The initial stage of flushing (the opening of buds and the exposure of new needles) on June 20 indicates a very short growing season (Figure 8). Cri Lake (Great Whale River Area) Site, Québec The Cri Lake site (Figure 4) is situated northeast of Great Whale River and a short distance inland from the coast of Hudson Bay (Figure 9). Increment cores were collected at this site on June 24, 1979. The general area of Great Whale River is in the middle of the subarctic forest, which is characterized by sporadic open stands of small conifers, chiefly black spruce, with lichens and mosses on the largely exposed granite bedrock (Figure 10). The importance of favourable microclimate and its effect on tree growth, especially in this area, is well demonstrated by the Cri Lake site. A grove of white spruce (a rarity from Matagami to Richmond Gulf) and adjacent black spruce trees, growing on well-drained soil ina depression with west-south-westerly aspect was located with the help of Roxanne Lajeunesse of the Centre d'Etudes Nordiques, Québec City. The site is well sheltered from strong winds by the surrounding hills, and the aspect is favourable for incoming solar radiation. Between Great Whale River and Cri Lake there are a number of hill crests of granite, some 150- to 200-m high. In general, the hills are aligned in an east-west direction, with streams draining into Hudson Bay. As a result of glacial action and the following marine phase, which washed out the finer materials from the moraines, bedrock surfaces were left strewn with fields of boulders and isolated larger blocks of rock. The terrain is irregular and open, approximately 90% without trees or shrubs. Some bedrock is covered with lichens and mosses. Small stunted trees were encountered only in low-lying depressions and stream beds. At the time the site was visited, remnants of heavy snow accumulation was still evident in areas with northern exposure (Figure 11). Sixteen white spruce and 15 black spruce trees were sampled with a Swedish increment borer (Figure 12). Two cores were collected from each tree. 143 æ d à Ca i=) à é Cri Lake FIGURE 9: Map of the Great Whale River (Poste-de-la-Baleine), Québec area showing the location of the Cri Lake tree-ring site in centre. 144 peut; : Aa enr fee ju Gt "rs * FIGURE 10: Cri Lake in the foreground with Hudson Bay and Bill of Portland Island near the horizon. 145 FIGURE 11: Remmants of snow near Crt Lake on June 24, 1979. 146 EU : co eo Sampling a Cri Lake white spruce tree with a Swedish increment borer. FIGURE 12: Limbs near the base of the tree were removed to facilitate sampling. 147 FIGURE 13: White spruce trees at Crt Lake with full crouns to ground level. Note the open, boulder-covered gromd. 148 White spruce trees were generally sound with an average breast-height diameter of 42.4 cm, ranging from 29.7 to 52.6 cm. The trees are growing in a pure, well-defined grove (a pocket with no other species), with heavy, live branches often close to the ground level (Figure 13). Trees were invariably upright, without compression wood and approximately 14-m tall. The oldest tree sampled was 297 years of age at breast height. Shorter black spruce trees were growing on lower and wetter ground to the northwest of the white spruce stand. Average diameter at breast height for the 15 black spruce sample trees was 19 cm, ranging from 16 to 21 cm, and the average tree height was approximately 9 m. Almost all of the samples were remarkably free of compression wood, indicating stable ground conditions that allow tree stems to grow in an upright position. The oldest black spruce sample tree was approximately 175 years of age. The black spruce trees were growing in an open stand with well-shaped, full crowns and with live branch whorls often extending down to the base of the stems. Ground cover was luxuriant, mainly mosses, lichens and crowberry. The ground was wet but not fully saturated. Richmond Gulf Site, Québec The most northerly tree-ring samples were collected at Richmond Gulf during the June, 1979, INQUA excursion (Figures 4, 14). The site is located near the tundra and close to the northern limit of tree growth. Trees were uncommon in the area; lichen, shrubs, grasses and sedges mainly represented the plant kingdom. Three small groves of black spruce, growing in association with birch, alder and willow shrubs, were found and sampled. All of the trees were stunted in growth with crowns shaped by strong winds. The underlying permafrost resulted in a shallow root zone. This factor, the harsh climatic conditions and the short growing season (the black spruce trees were just beginning to flush on June 23, the date of collection) all combine to limit tree growth. Black spruce trees were all dwarfed to a shrub-like form, approximately 2.5 to 3.3 m tall, with an average diameter of 8.9 cm, ranging from 6.1 to 14.0 cm. The oldest trees were approximately 230 years old, with 36 annual rings per cm. A majority of the trees were reproducing by layering (i.e. the lower live branches, on becoming covered by mosses and litter, contacted the soil and developed roots and eventually became trees). Trees were sampled from three subsites, two of which (subsites a and b) had northeasterly aspect, were 149 150 FIGURE 14: © SE + _- Richmond Gulf tree-ring site near the northern treeline, Québec. photographs are taken from subsite c. Substte b ts outlined in the rectangle. Scale can be judged by the three persons (circles) shown in the photographs. Ratsed beaches, which appear as horizontal lines on the ground surface, have resulted from crustal uplift after deglaciation. The LST gently sloping, and were approximately 90 m above sea level. The third subsite (subsite c) was located at the same elevation, but north of the other two subsites, on a small plateau on top of a rock bluff that sloped to the west. Twenty-six living trees were sampled with a hand saw. Complete cross-sectional disks were taken because of the pronounced eccentricity of the stems and tight ring growth (Figure 15). These site descriptions are presented here to indicate the diversity of tree-ring sites in the Boreal Forest Region in northern Québec. As they become available after processing with an x-ray densitometry system, tree-ring data will be presented in subsequent publications. Part 2 (following) is the first of these reports. 152 FIGURE 15: Wature of tree-ring growth at Richmond Gulf. Note the high-density compression wood and off-centre pith. Coin is for: scale. 153 PART 2: WHITE SPRUCE ANNUAL RING WIDTH AND DENSITY CHRONOLOGIES FROM NEAR GREAT WHALE RIVER (CRI LAKE), QUEBEC M.L. Parker, L.A. Jozsa, Sandra G. Johnson and Paul A. Bramhall BACKGROUND Tree-ring width and density chronologies are the end products of a complex array of factors related to the environment (including climate, site condition and geographic location) and also to the genetic nature of individual trees and tree species. Adding to the complexity is the fact that a number of different techniques can be employed to make these data useable. The major objective in collecting, processing and presenting the white spruce tree-ring chronologies from the Cri Lake site is to provide summary chronologies that can be compared with other forms of proxy data for purposes of reconstructing past climate. Corollary objectives are to evaluate the dendrochronological quality of “ice spruce from the area and to produce tree-ring chronologies that can be used for dating purposes. The raw-ring width and density data comprising the tree-ring series contain several components. To make this preparation of the data most useful for the stated purposes, we will define three components that seem to be present in all of the tree-ring series that we have observed through the years. These components are: (1) growth trend from pith to bark (the "A" component); (2) short-term fluctuations, greater than 10 years in length and shorter than the ring series of the tree being studied (the "B" component); and (3) year-to-year fluctuations, with all fluctuations greater than 10 years removed from the series (the "C" component). The growth trend for a given species is mainly a function of the age of the tree and is essentially non-climatic. However, Mitchell (1967) found that the general shape of the growth trend for black spruce varied with latitude, being flatter for trees growing at more northerly locations. The growth trend, however, is not affected by year-to-year variations in climate. 154 In good dendrochronological-quality tree-ring series, short-term fluctuations and year-to-year fluctuations are mainly a function of direct climatic factors, such as temperature and precipitation. Some other factors, such as fire, insect damage and competition between trees, also affect the nature of the short-term and year-to-year fluctuations. These "other" factors also may be indirectly related to climate, but they generally lead to interpretation complications, and an effort should be made to minimize these effects when building a "climatic" tree-ring chronology. The two tree-ring parameters selected to be presented for the Cri Lake trees are ring width and maximum ring density. These two variables have been found to be closely related to climatic factors in previous studies. Somes examples of raw data for an individual increment core will be given, and then summary chronologies for both ring width and maximum ring density of the three numbered components above will be presented. Information is given about the site and nature of the collection and about processing techniques used to obtain the data. THE CRI LAKE SITE Cri Lake is located 4.2 km northeast of Great Whale River, Québec, and 2.5 km from the shore of Hudson Bay. The forest in this area of Canada is dominated by black spruce, white spruce being quite rare. However, an isolated stand of white spruce grows around the edge of Cri Lake. Black spruce trees occur nearby, but there is little overlapping of the species in the stands. Apparently drainage and other conditions are proper for the growth of white spruce at this location. Increment cores were taken from both white spruce and black spruce trees. The white spruce are growing mainly on the southeast, east and northeast sides of the lake, and the black spruce are growing on the west and northwest sides. The white spruce are generally older than the black spruce: the mean age of the white spruce trees, from which cores were collected, is about 250 years and the black spruce mean age is only 150 years. Usually, white spruce is of better dendrochronological quality than black spruce; therefore, the white spruce samples were selected as the first to be processed on this project. Tree statistics of the white spruce cores collected at Cri Lake are given in Table 1. 155 TABLE 1: DIAMETER AT TREE BREAST HEIGHT (CM) 1 52.6 2 35.6 3 46.0 4 38.6 5 45.7 6 GRA {| 7 45.7 8 41.1 9 41.9 * ** Rings present but too rotten to measure. 156 Compass direction. w* = Z NE TREE STATISTICS OF WHITE SPRUCE AT CRI LAKE, QUEBEC. NUMBER OF RINGS AT BREAST HEIGHT COMMENTS Rotten stand Rotten Rotten centre; centre** centre** on edge of In collecting tree-ring samples for climatic studies or for building chronologies for dating purposes, an effort is made to collect from the oldest and most climatically sensitive trees on the site. Usually, samples are taken from about 15 trees with the intention of processing about 10 samples. Cores from some trees are rejected because of rotten segments, knots, damage by the sample-preparation saw, or for any other reason that may reduce the quality of the sample for analysis. Of 16 trees sampled at Cri Lake, nine were used in building tree-ring chronologies, the others being rejected for reasons mentioned above. PROCESSING TECHNIQUES Ring-width and ring-density data were derived from the Cri Lake white spruce increment cores by using the technique of x-ray densitometry. This method is relatively new to the field of dendrochronology and was pioneered in France by Hubert Polge (1963, 1966). The techniques and instruments of x-ray densitometry used at the Western Forest Products Laboratory have been described in detail (Parker 1970, Parker and Jozsa 1973, Parker et. NO TS 1077/1980) DUE mR theya ae br ie (1) increment cores are air dried and mounted with glue between two mounting sticks; (2) mounted cores are cut to a uniform thickness on a twin-blade saw built for this purpose (Kusec 1972); (3) samples are labelled with x-ray opaque paint; (4) a portion of one of the core mounting sticks is cut away and the cell angle (the angle formed by the long axis of the longitudinal tracheid cells and the long axis of the increment core) is measured under a microscope; (5) prepared cores are placed on a sheet of x-ray film along with calibration wedges and exposed at the proper angle by a moving x-ray machine; (6) x-ray film is developed under carefully controlled conditions; (7) the radiograph is then examined on a light table under a low-power microscope and calendar-year dates and other information are marked on it; (8) radiographs are then scanned on a computerized densitometer that converts the image of the increment core into ring-width and ring-density data; (9) data are stored on magnetic tape for further processing and summarizing. 152 CRI LAKE WIDTH AND DENSITY DATA PRODUCED BY THE X-RAY DENSITOMETRY SYSTEM At the Western Forest Products Laboratory, tree-ring data are produced and processed on a computerized densitometer (Parker et al. 1973) and on our mini-computers. Data acquisition and processing programs have been described (Parker et al. 1980). The major data acquisition program, Tree-Ring Input Program (TRIP), controls the densitometer and computer in the conversion of the images on the x-ray film to width and density data. Appendices 1 and 2 give examples of the form of data stored on magnetic tape by TRIP. Appendix 1 presents data for all of the annual rings scanned by the densitometer on one increment core from the Cri Lake samples. For each annual ring, the table in this appendix shows the year, the distance from the pith and the intra-ring width and density variables: (1) ring width, (2) earlywood width, (3) latewood width, (4) ring density, (5) earlywood density, (6) latewood density, (7) minimum ring density, and (8) maximum ring density. Note that for some years there is a zero value for latewood width and latewood density. A certain density level (.45 g/cm3, oven-dry weight and volume) was used to define the earlywood-latewood boundary. Some rings do not attain this level of density. If a zero value is introduced for latewood width or latewood density, the average for these parameters is disproportionately influenced. Therefore, if the ring series contain zero-latewood annual rings, it is better not to use latewood width or latewood density for climatic studies. Ring width and maximum ring density, the two variables used in the Cri Lake analysis, are not affected by this zero-latewood condition. The other table that presents increment core data stored by TRIP, Appendix 2, shows data for only three rings (1799-1801). In addition to the intra-ring width and density variables, 100-value intra-ring density profiles also are shown in plotted form, first in the standardized form with 100 units per ring, and then the three rings are shown after ring width has been reintroduced into the intra-ring density profile (Figure 16). The ring width and maximum ring density summary chronologies for Cri Lake were compiled from data obtained from 18 increment cores, two cores from each of nine trees. Appendix 3 gives the mean, standard deviation, and mean sensitivity of the intra-ring width and density variables for these cores. Mean sensitivity (Schulman 1956) is a measure of year-to-year variability (the higher the value, the greater the variation). 158 1801 Density (g/cm?) | | [+ 100 +4 ——100 "Î 100 "+ Standardized — Width Units 0.7 0.6 1800 1801 0.5 0.4 Density (g/cm?) 0.3 1 1 1 ———+}-——0.85—+}- 0.7 —+t- 1 : 4e+——-—-1.22 ‘ Actual Ring Width (mm) FIGURE 16: Examples of intra-ring density profiles of three annual rings (1799-1801), with the earlywood to the left and latewood to the right. The three rings are shown first in standardized form, t.e., 100 units per ring, and then they are shown in actual ring width form (mm) (see Appendix 2). 159 Total ring width and maximum ring density are the variables that have been selected for processing of the Cri Lake trees. However, some of the other intra-ring width and density variables, or the intra-ring density profiles, may prove to be useful in providing additional information about past climatic conditions. STANDARDIZING AND SUMMARIZING TECHNIQUES The "raw data" used for the Cri Lake site consist of: (1) ring-width values in 0.01 mm units, (2) maximum ring-density values in g/cm3 measured at 0.01 mm increments along the scanned portion of an increment core. These raw data are "standardized" by converting them to indices (ratio of observed value to fitted trend, yielding values with a mean of 1.00) by removing the growth trend (and sometimes other trends or fluctuations). The standardized data (indices) for all cores are then "summarized", or averaged, to produce a summary or "master" chronology for the site. This procedure is illustrated in Figure 17. In a previous section of this publication, the "A", "B" and "C" components were mentioned: "A" being defined as the growth trend; "B", the short-term fluctuations greater than 10 years in length; and "C", the year-to-year variations. Various processing programs are used to produce an "A", a "B", a "C" and a "B & Cc" chronology from the raw data of each increment core. These data are then averaged with data from other cores to produce the summary chronologies described in the next section. If increment core data are averaged without removing the growth trend, non-climatic fluctuations related to the age of the tree will be included in the chronology. Therefore, it is essential that the growth trend be removed. However, variation due to climate may be removed inadvertently, if proper techniques are not used in this procedure. A method for removing the growth trend from tree-ring series, that was applied to the Cri Lake data, was developed by Warren (submitted 1980). This method is designed to remove the growth trend and the non-climatic surges in growth, such as tree growth release after a forest fire. Tree-ring indices (in the form of our "B & C" chronology) are produced by calculating deviations from a growth-trend line through the raw data values. Warren's method, or the "“compound-increment-function (CIF)" method, is designed so that units of data series (of predetermined length) may be tested for goodness of fit, as the "growth trend" or estimated curve is being calculated for the entire ring series. The length of the unit 160 Raw Ring Width Data Raw Data and € £ = Q 2 = ru) z ia 1700 1750 1800 1850 1900 1950 2000 FIGURE 17: YEARS Components of a ring width data sertes of an inerement core from a Cri Lake white spruce tree (Tree 1, Core a). The growth trend or "A" component ts superimposed on the broken-line plot of the raw ring width data. The "B" & C" component ts the chronology that remains after the "A" component has been removed. Short-term fluctuations are represented by the "B" component and the "C" component conststs of the year-to-year fluctuations. 161 selected to be tested for goodness of fit determines the duration and magnitude of the fluctuations that will be removed from the series. The entire length of the ring series was selected to be used as the test unit for the Cri Lake cores. This was done to assure that only the age trend would be removed and that mo short-term, climatically-related fluctuations would be removed. Non-climatic fluctuations may be left in the series by this approach, but these can be removed in a subsequent step. Applying the CIF method in this way is not using it exactly as designed, but the desired results were obtained by using it as described. The "B" and "C" chronologies are obtained from the tree-ring indices (the "B & C" chronology). This is done by using a digital-filter technique (Parker 1970). The "B" chronology is the estimated curve which is a weighted running mean. Deviations from this line are the year-to-year fluctuations, or the "C" chronology. RING WIDTH AND MAXIMUM RING DENSITY SUMMARY CHRONOLOGIES FROM CRI LAKE SAMPLES Summary chronologies are obtained from the ring width and maximum ring density data by removing the growth trend and averaging the values for the 18 increment cores into one composite ring series for each parameter (ring width and maximum ring density) and for each component ("A", "B", "C", and "B & C"). These summary chronologies are presented in Figures 18 to 21 and Appendices 4 to 9. The "A" component chronologies (the ring width growth trend and the maximum ring density growth trend) are constructed by averaging the estimated curve values (obtained by the CIF technique) of all the increment cores. These chronologies are averaged by age from pith rather than by calendar year dates of the rings; i.e., the pith date is left justified for each ring series before all of the series are averaged together to produce one growth trend series, which is an average of all the growth trends of all the cores. These trend lines were determined mathematically and then slightly smoothed by a hand-drawn line. Note that there is a marked and regular downward curvilinear trend in the ring-width growth trend component (Figure 18). Of the 18 cores, 11 have a downward trend in maximum ring density, four have an upward trend and three are virtually neutral in trend. The range in density for the maximum ring density composite series is 0.65 g/cm3 at the pith end and 0.57 g/cm3 at the bark end. 162 FIGURE 18: INDICES un Wu © Q < INDICES 1700 1750 1800 1850 1900 1950 1977 YEARS Ring width and maximum ring density summary chronologies for the nine Cri Lake white spruce trees. Represented are: (1) growth trend ("A"); (2) a combination of short-term and year-to- year fluctuations ("B & C"); (3) short-term fluctuations ("B"); and (4) year-to- -year fluctuations cron) (Appendices 4-9). Data are given on correlation coefficient (r) and number tn sample (n). 163 RING-WIDTH AND MAXIMUM-DENSITY INDICES ("B & C”) 1922 164 FIGURE 19: Ring Width “B & C” Maximum Ring Density ‘’B & C” 1952 Cri Lake white spruce ring width and maximum ring density summary chronologies, "B & C" components (Appendices 4, 5). 2822 RING-WIDTH AND MAXIMUM-DENSITY INDICES ("B”) FIGURE 20: Ring Width “B”’ 1958 Cri Lake white spruce ring width and maximum ring density summary chronologies, "B" component (Appendices 6, 7). 165 2028 RING-WIDTH AND MAXIMUM-DENSITY INDICES (“C”) FIGURE 21: ra Ring Width “‘C” p bo NP HN | Ass ri Nf abl } I \i V \ Maximum Ring Density “C” 1958 Cri Lake white spruce ring width and maximum ring density summary chronologies, "C" component (Appendices 8, 9). 2828 The tree-ring indices that comprise the "B & C" components chronology (a combination of the short-term and year-to-year fluctuations) are obtained by averaging the deviations from the growth trends of all series for the 18 cores on a calendar-year basis. A correlation coefficient (r) of 0.30 (number in sample (N) = 278) was obtained from a comparison of the ring width "B & C" chronology and the maximum ring density "B & C" chronology (Figure 19). The "B & C" summary chronologies were then run through a digital filter to obtain the "B" summary chronology (short-term fluctuations) and the "C" summary chronology (year-to-year fluctuations) (Figure 18). This particular digital filter derives the estimated curve (weighted running mean) by considering a central value and six values on each side of this central value. The values nearer the central value are given the most weight. Correlation values are: r = 0.48 for the "B" chronology comparison of ring width and maximum ring density (Figure 20), and r = 0.22 for the "C" chronology comparison (Figure 21). These correlations indicate that ring width and maximum ring density are not responding in exactly the same way to climatic factors. One way to get a general impression of the dating quality of trees at a given site is to observe how well the ring series from one tree compares with the site composite chronologies. This was done for the Cri Lake material by comparing the ring width and maximum ring density series of a core from Tree 1 (selected at random) with the summary chronologies. (Although the core compared was used to build the chronologies, it has limited influence on the summary chronologies because each of these summaries was built from 18 ring series). The results of these comparisons are presented in Figures 22, 23 and 24. The correlation coefficient values obtained from comparisons of this individual core with the summary series are: r= 0.90 (n = 244) for a comparison of the "B" component of ring-width chronologies; r = 0.83 for a comparison of the "B" component of the maximum ring density chronologies; r = 0.80 for a comparison of the "C" component of the ring width series; and r = 0.90 for the "C" component, maximum ring density comparison. In order to make an additional check on the validity of these high correlations, it was decided to make the comparisons again for the period all cores have in common: 1910 to 1975 (n = 66). The results of these comparisons of the individual core with the summary series 167 SHORT-TERM FLUCTUATIONS (“B”) un Lu 2 Q = = œ 1700 1750 1800 1850 1900 1950 2000 Site Summary Chronology MXD INDICES 1700 1750 1800 1850 1900 1950 2000 FIGURE 22: Comparison of Cri Lake site summary chronology (ring width (top) and maximum ring density (bottom)) with individual-core data, "B" component. Note high degree of correlation. 168 RING-WIDTH INDICES ("C7 YEAR-TO-YEAR FLUCTUATIONS (“C”’) Individual Core Chronology A LL oo 0... Le 1900 1950 2000 FIGURE 23: Comparison of Cri Lake ring width site summary chronology with individual-core chronology, "C" component. Note high degree of correlation. MAXIMUM-DENSITY INDICES ("C”) ae Core Chronology A) Al AA SAVE te AN [|| PL EM A MAI VE UMA AAA LME Site Summary Chronology NN MU A LA MORAL ANA AV AUS AURA PA AA Len mo D A FIGURE 24. Comparison of Cri Lake maximum ring density site summary chronology with an individual-core chronology, "C" component. Note high degree of correlation. 1950 (1) ring width , "B" component (r = 0.73); (2) maximum ring density, "B" component (r = 0.66); (3) ring width, "C" component (r = 0.78); 0.88). (4) maximum ring density, "C" component (r The similarity of the individual-core chronologies with the summary chronologies is an indication that the Cri Lake trees are of good quality for dating purposes and for using to reconstruct climatic history for the Great Whale River area. CONCLUSION The general objectives in conducting this research have been to locate and collect good quality tree-ring samples from the James Bay-Hudson Bay region and to produce tree-ring chronologies that can be used to reconstruct climatic history and for other purposes, such as dating or assessing forest growth. On the basis of visual comparisons, the Cri Lake white spruce trees seem to provide the best quality chronologies for the area, but this will be determined in more detail after the other sites are processed. It is noteworthy that the mean age of the white spruce collected at Cri Lake is about 250 years and the mean age is only 150 years for the black spruce growing at the same location. Probably some species-related differences are responsible for this, rather than attributing the age of the trees at the site to a chance factor such as fire. For the Cri Lake material, an effort has been made to establish a standard format for processing the tree-ring data that can be followed for processing material from other sites in the area. The new approach of breaking the chronologies down into components and producing summary "A", "B", "C", and "B & C" chronologies may prove useful in providing more information about the relationships between tree-ring data and climatic factors. For example, the "Short-term" fluctuations ("B" chronology) may relate to climate determined by one set of causal factors and the year-to-year fluctuations ("C" chronology) may relate to climate determined by another set of causal factors. The good correlations obtained, when comparing individual core chronologies with summary chronologies, indicate that the tree rings are reflecting some general environmental conditions (climate) and not just some factors that influence the growth of individual trees. 7 In most tree-ring studies, only total ring width has been analyzed. Breaking the annual ring into earlywood and latewood components, and using density as well as ring width, is an approach that provides considerably more information that can be related to climate. This procedure leads to a more detailed and accurate reconstruction of climatic history. Two good-quality parameters, ring width and maximum ring density, were processed to produce the Cri Lake summary chronologies. However, other data, such as minimum ring density, latewood width, etc., have been measured and stored on magnetic tape and are available for future analysis. Now that the tree-ring summary chronologies are built for Cri Lake, future research will involve: (1) relating these data to instrumental weather records and to proxy historical climatic data, and (2) processing tree-ring data from other sites in the James Bay-Hudson Bay region. ACKNOWLEDGEMENTS We would like to thank Dr. C.R. Harington, Paleobiology Division, Museum of Natural Sciences, National Museums of Canada for his organization, encouragement and financial support of this project. We also thank Dr. Claude Hillaire-Marcel, Département des Sciences de la Terre, Université du Québec à Montréal, and other organizers and participants of the INQUA Hudson Bay Field Meeting, for their assistance in the field work. Also, we appreciate the help of Roxanne Lajeunesse in locating the Cri Lake site and for assisting in the collection of tree-ring samples. We thank Dr. W.G. Warren for developing the compound-increment-function growth-trend- removing technique. This research was supported in part by the Canadian Forestry Service, Environment Canada, under contract DSS Legal Agreement No. 05SS KL229-9-4196. 172 REFERENCES Kusec, D.J. 1972. Twin-blade saw for precision machining of increment cores. Wood and Fiber 4(1): 44-49. Mitchell, V.L. 1967. An investigation of certain aspects of tree growth rates in relation to climate in the central Canadian Boreal Forest. Env. Canda, Contract OS04-0206. Parker, M.L. 1970. Dendrochronological techniques used by the Geological Survey of Canada. In: Tree-Ring Analysis with Special Reference to Northwest America. Edited by: JeHeGe Smith and J. Worrall. Univ. British Columbia, Fac. Forestry Bull. No. 7: 55-66. (also published in 1971 as Geol. Surv. Can. Paper 71-25: 1-30). - 1976. Improving tree-ring dating in northern Canada by x-ray densitometry. Syesis 9: 163-172. Parker, M.L., R.D. Bruce, and L.A. Jozsa. 1977. Calibration, data acquisition and processing procedures used with an on-line tree-ring scanning densitometer. Int. Union Forest Res. Organ., Group P4.01.05 Meeting, September, 1977, Corvallis, Oregon. 20 pp. - 1980. X-ray densitometry of tree rings at the WFPL. Forintek Can. Corp., Western Forest Prod. Lab., Tech. Rep. No. 10: 1-18. (this publication is a modification of the above report). Parker, M.L., and L.A. Jozsa. 1973. X-ray scanning machine for tree-ring width and density analysis. Wood and Fiber 5(3): 192-197. Parker, M.L., and R.W. Kennedy. 1973. The status of radiation densitometry for measurement of wood specific gravity. Proc. Int. Union Forest Res. Organ., Div. 5 Meetings, September-October, 1973, Cape Town and Pretoria, South Africa. pp. 882-893. Parker, M.L., Je Schoorlemmer, and L.J. Carver. 1973. A computerized scanning densitometer for automatic recording of tree-ring width and density data from x-ray negatives. Wood and Fiber 5(3): 237-248. Polge, H. 1963. L'analyse densitométrique de clichés radiographiques: une nouvelle méthode de détermination de la texture du bois. Ann. Ec. Natl. Eaux Forêts, St. Rech. Expér. 20(4): 530-581. - 1966. Etablissement des courbes de variation de la densité du bois par exploration densitométrique de radiographies d'échantillons préléves à la tarière sur des arbres vivants. Ann. Sci. For., Nancy 23(1): 1-206. Rowe, J.S. 1972. Forest regions of Canada. Env. Can., Can. Forestry Serv., Publ. No. 1300, Ottawa. 171 pp. Schulman, E. 1956. Dendroclimatic changes in semiarid America. Univ. Ariz. Press, Tucson. 142 pp. Warren, W.G. (submitted 1980). On removing the growth trend from dendrochronological data. Tree-Ring Bull. 1573 APPENDIX 1: DATA ON ALL OF THE ANNUAL RINGS (1732-1975) FROM A SINGLE CRI LAKE WHITE SPRUCE CORE (STORED ON TAPE BY THE TREE-RING INPUT PROGRAM (TRIP) OF THE WESTERN FOREST PRODUCTS LABORATORY). YEARS: 4732-41975 YEAR RSX RW * EU* DR RTC CET" LD** MND** MXD* * 4732 4004 444 33 27 4044 3450 S725 2861 6593 1733 4442 160 123 37 4200 3644 6044 3177 8596 1734 4272 432 405 C7 05814 3306 S739 2763 7640 4735 41404 108 58 50 4686 3388 6187 3005 8902 4736 41542 472 108 65 4596 3422 6552 2644 9048 4737 4684 114 87 27 3990 3367 S948 2882 7697 1738 1798 80 54 26 4442 3692 S879 3094 7354 4739 1878 97 74 22 "3469 3420 “Sapa 2532 67412 4740 4975 94 78 1 3404. (S224 4475 2789 | 5083 1741 2066 404 77 Sa 8779 | S255 SS2i. 2731 6210 4742 2470 402 83 19 3574 3492 S267 2618 6060 1743 2272 97 67 ao 5957 3354 S349 2874 “4622 4744 2369 80 57 2% 4044 3403 548? 2988 6860 4745 2449 405 85 Al 3732 $238 “Se7e 20906 yosa 1746 2554 409 96 43 3234 3001 4965 2S2a ‘S$sS6 4747 2663 93 78 45 3488 3160 S238 2498 6292 4748 2756 87 64 2 3844 3068 S759 2500 7445 1749 2843 95 72 22 3798 3455 S869 2644 7429 1750 2938 84 73 44 3479 3307 4648 2684 S232 1751 3022 429 95 34 3920 3170 S997 2679 8196 1752 34Si 449 86 $3 3035 3136 S644 2472 6430 1762 3270 86 67 20 3494 3079 4893 2644 5544 $754 33456 88 73 45 3513 3126 $448 2579 4572 4755 3444 74 62 40 3258 3084 4360 26416 4689 4756 3545 74 58 14 3734 3593 S155 2958 6106 1757 3586 73 73 6 3447 ‘3147 0 2644 4434 1758 3659 60 53 7 3343 3198 4278 2582 4739 1759 3749 S6 47 9 3442 2865 4429 2528 4927 4760 3775 70 63 8 3425 2948 4547 2466 4903 1761 3845 Si 42 9 3266 2954 4722 “2500 “S270 4762 3896 63 53 40 3464 2883 4634 2439 5257 4763 3959 69 5s 45 3529 2984 S526 2495 6894 4764 4028 60 43 i7 x24 3223 S345 “2585 6435 4765 4088 87 82 2832 2754 4480 2299 4541 5) = 3746 3163 5679 2540 6611 1767 4282 93 84 9 3120 2947 4752 2386 $267 4 $292 2884 $433 2506 6196 0 3558 2874 5610 2427 6823 * Width values in .04 mm units -- RSX = distance from pith; RW = ring width; EW = earlywood width; LW = latewood width. ** Density values in .0004 g/cm? units -- RD = ring density; ED = earlywood density; LD = latewood density; MND = minimum density; MXD = maximum density. 174 APPENDIX 1: YEAR (cont'd) 4624 4716 4778 4825 4892 A4 5007 5060 Se 54158 S221 5289 5357 5431 S489 S555 5632 5701 5785 5894 5985 6097 62it 6304. 6395 6500 6643 6742 6853 6922 7044 74129 7205 7307 7375 7499 7601 7705 7830 F927 8042 8104 8190 8284 8369 8496 8640 8666 8723 8759 8817 8862 8912 8955 8981 LU RD * 3276 3131 3698 3105 3220 3189 3348 3108 AT 3154 3191 3838 3737 3419 AVE 3477 3415 3244 3547 3882 3242 3397 3380 3536 3589 3307 2983 3335 3497 3542 3159 3770 ip hah 33714. 3324 3485 3149 31416 3568 3246 3293 3121 3032 3344 3107 3440 3349 2859 3709 3295 3434 3425 3070 3453 3835 Rare) Sida 2934 2955 2914 3170 34116 2916 3084 3049 3057 2826 2756 2899 3066 2994 3016 3172 2931 3068 2838 2932 2907 2978 3070 2996 2812 2933 2840 3063 3050 2969 3039 2859 3047 3295 3077 3040 2898 3067 3420 * ET 5459 41,97 S02? 4262 4776 490 4. 5165 4589 5552 5064 4649 9873 Soe 5494 it) Dre 5304 A701 STE S939 4637 $273 4754 5739 5594 6124 5004 5750 5626 5532 4366 5564 5209 Sere 5996 5496 4847 4413 5267 49:4 5760 S082 4898 S051 4264 $255 4778 0 5693 0 4861 4985 4375 4504 4354 MND 2585 2644. 2848 2017 2448 2564 2646 2604 2680 2527 2472 2585 AAS EAS! 2528 2585 2528 2585 2306 2504 2641 2489 2472 2640 2590 2590 2311 243% 2445 2364 2394 2635 2528 25e 2496 2306 2378 OK 2306 2528 2439 224 2306 2197 2364. 2584 2417 2641. 2306 2695 2870 2489 2750 2495 2778 2947 MX D ** 6578 4996 5760 4773 5604 5454 8315 5312 7050 5954 520% 7734 8060 6796 4342 6256 6092 5048 7484 7925 5209 6786 S571 7428 6706 8032 5968 7449 7050 6773 4736 7001 6254 6255 7842 7545 5665 4916 5981 5976 7877 6136 $855 64.72 AST7A 6181 S744 3931 7052 44314 S757 6040 4962 S271 4839 175 APPENDIX 1: 176 YEAR 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1935 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 (cont'd) 8999 9013 9047 9075 9106 9135 9169 9200 Dent 92714 9313 9337 9379 9404 9430 9460 9488 9548 9547 9566 9575 9682 9657 9694 9789 9763 9799 9839 9881 9930 9980 10029 10083 40429 10180 40248 10257 10312 10374 10431 10494 10560 10612 108659 10716 10749 10794 10835 10885 10928 10988 11039 11095 11149 11202 Lu > > » => E INNVUILSS NOUS NNOLNE Tome hore omauiuonmancvouvss eh RD** 4325 3533 3421 4007 4059 3671 3750 3381 3038 3093 3407 2903 3313 3159 3098 3624 3333 3459 3905 3695 3281 3450 3150 3207 3449 3256 3341 3543 2937 3349 3141 3421 3394 3763 Doi 3489 3200 3305 3493 3335 3364 3826 3502 3098 3376 3634 3189 3275 3644 3273 3784 3529 3564 3729 3465 4517 4821 4487 $252 25e 4862 5080 4400 0 4704 423: 4054 4566 0 4477 4966 4691 0 4170 $4134 0 5409 4432 4638 5189 4724 47 $8 5051 0 4570 4614 5408 4924 5324 4693 4894 4768 0 4739 5045 4739 5247 4956 4476 4542 5262 4175 S054 S159 5022 5897 5699 5503 5955 5592 * MND’ CAES 2825 2928 2988 3214 2970 3035 2918 2472 2411 2698 2361 2618 2665 2328 2687 2696 3045 3461 PAS 2865 2488 2461 2573 2614 2641 2801 2552 2528 2687 2528 2528 2746 2795 2824 2623 2692 2755 2868 2732 2664 2993 2865 2634 2842 2763 2698 2466 2528 2585 2641 2528 2696 2547 2612 ek MXD 5257 5805 4923 6255 6386 5454 5979 4764 4378 5523 4716 4504 5084 4332 5432 5840 5277 4227 4565 6165 4465 6859 5035 5533 6208 5529 5596 6092 4088 5092 5207 6877 5720 6443 5197 5740 5447 4418 5532 6032 5389 6330 5900 4988 4999 6620 4577 6032 6285 5827 8064 7220 6747 7654 7402 APPENDIX 1: (cont'd) YEAR RSX* RW * Ew* we RD * ED* LISA AND eorxp= 1980 411249 49 44 GS 3028 2898 4475 2506 4699 1881 11298 54 40 14 3684 ?842 6093 2266 £7868 4882 44352 S7 48 9 3338 2983 S228 2495 6404 4883 414409 58 49 9 3548 3302 4886 2770 5589 4884 44467 63 57 é 2969 2818 4404 2252 4804 1885 44530 69 63 b 3268 3439 4643 2687 4964 4886 441599 66 55 44 3538 3064 S906 2447 7574 4887 44665 67 S7 40 3275 2937 S?P03 2523 6262 1888 11732 64 57 4 2948 2864 4450 2519 4586 4889 414793 53 42 44 3568 3105 5336 2687 6834 1890 411846 53 46 7 3495 2958 4749 2482 54698 4894 411899 S0 4b 4 3097 #3000 4208 2528 4547 1892 44949 Ab 36 40 3473 2939 5397 2444 6857 1893 11995 58 49 9 3242 2802 5449 2450 6931 4894 412053 52 47 GS P8814 2729 4317. 2377. 4782 4895 42105 64 49 12 3329 2776 S583 2295 7006 1996 124166 Ab 37 9 3527 3087 S337 2641 65414 1997 42242 4S 32 13 3847 3463 5529 2757 6978 1898 42257 44 32 9 3584 3492 4980 2692 5667 1899 12298 46 34 42 3797 3209 $463 2750 6818 4900 412344 38 34 4. S377 S206 4603 2739 5004 4904 412382 56 45 44 3442 2987 5303 2517 6616 4902 412438 45 38 7 3493 2935 4594 2540 5330 1903 12483 56 48 8 3196 2922 4842 2364 5540 4904 412539 45 33 42 3816 3398 4964 3059 £5742 4905 412584 45 38 7 3446 3194 4624 3046 S324 1906 412629 49 40 9 3549 3486 S159 2898 6181 1907 12678 50 42 8 3624 3356 5028 2974 5886 4908 412728 42 34 8 3447 34143 4579 2747 5208 4909 42770 39 29 40 3468 2967 4922 2490 5583 4940 412809 43 37 44 3705 3280 5136 2864 5988 4944 412857 35 24 44 3780 3106 $250 2874 6405 4942 12892 30 30 0 3444 3444 0 2954 4462 1913 412922 35 27 8 3443 3009 4779 2755 5814 4944 42957 4. 34 7 3256 2963 4677 42584 5279 4945 412999 40 34 9 3495 3003 5190 2695 6240 1916 13038 49 39 10 3666 3442 S823 2585 7657 4947 13087 S7 45 42 3612 34100 5534 2603 6925 4918 13144 40 34 9 3625 3274 4835 2866 5376 4949 413184 49 35 14 3999 3267 S825 2844 7734 1920 43233 37 27 40 4030 3556 $3410 3484 6525 4924 13270 52 42 40 3548 3456 5193 2698 61646 4922 413322 42 34 44 3825 3378 5085 3026 5944 4923 413364 64 59 GS 3418 3046 4320 2670 4653 1924 143428 64 S4 43 3573 3053 S613 2640 7480 4925 413492 54 43 44 3429 3403 4705 2575 5375 4926 13546 36 26 40 3693 3249 4925 2807 6033 1927 413582 36 26 40 3874 3174 S683 28614 7029 4928 43618 46 34 42 3738 3445 5420 2630 6826 4929 13664 34 25 9 3855 3496 4854 3166 5523 1930 413698 44, 32 9 3470 3027 S043 2740 5966 4934 13739 42 35 7 3494 2908 4623 2644 5408 4932 413784 38 34 8 3450 3004 5491 2585 6400 1933 13819 36 30 7 3477 2812 4738 2415 SSii 1934 43855 32 32 0 29142 2942 0 2428 «3972 dr. APPENDIX 1: 178 YEAR 1935 1936 L937 1938 1939 1940 1941 1942 1943 1944 4945 1946 1947 1948 1949 1950 1951 RAT TE 1953 1954 1955 1956 1957 1958 1959 1760 1964 1962 1963 1964 1965 1966 1967 1968 1969 1270 1971 1972 1973 1974 1975 (cont'd) RSX* RUW* Ew * LU 13887 25 17 g 43942 iS 44 4 13927 26 18 8 13953 23 is 8 13976 30 24 6 14006 23 45 8 14029 34 27 8 14063 27 49 8 44090 26 21 5 14116 25 47 8 14141 18 ii 2 J 14159 28 28 0 14187 23 19 4 442410 32 25 8 14242 20 42 8 14262 44 7 7 14276 33 23 44 14309 40 29 12 14349 42 34 8 14391 32 20 12 14423 29 49 40 14452 44 44 0 44496 32 27 5 14528 44 37 2 44572 45 36 9 14617 36 29 7 14653 30 23 7 14683 32 25 7 44745 43 43 0 14758 23 45 8 14781 20 20 0 14804 32 23 9 14833 42 35 7 14875 40 30 10 44945 4i 34 7 14956 47 38 9 45003 35 25 40 15038 42 32 40 45080 36 26 10 15116 44 29 42 45457 44 28 43 MEAN: .58 .47 44 S.D.: 1 .29 . 24 .08 SENS : 49 26 .56 16,D, = Standard deviation 2MSENS = Mean sensitivity ED ** 3247 3395 5432 314% 3026 3340 3043 3293 2988 3223 3290 3340 3230 2964 3026 3543 3409 3457 3401 3335 3490 3154 3164 3407 3018 3197 3464 3183 2970 3364 3569 3298 2926 3285 3353 3235 33596 3147 3095 3212 3222 L.D** Se 22 4417 5258 5507 4834 4802 5015 5040 AS24 S141 4699 0 4204 5086 4831 4843 5127 5494 4955 5947 5815 0 4183 4429 S304 4475 5253 5063 0 SA 0 5149 4412 4760 4258 5487 5139 474 5633 5387 5077 APPENDIX 2: DATA FOR THREE RINGS (1799-1801) FROM A SINGLE CRI LAKE WHITE SPRUCE CORE, INCLUDING INTRA-RING DENSITY PROFILE VALUES (FIGURE 16). YEARS: 1799-1801 YEAR RSX* RU * EU * LU* RrD** ep** LD** wnp** mxp** 1799 = 6922 422 97 25 3542 .2994°° coco Sa94" 773 2698** 2626 2559 2435 2445 2466 2454 2562 2555 2596 2583 2596 2618 2656 2660 2604 2564 2544 2518 2479 2453 2479 2565 2596 2691. 2730 2767 2750 2781 2786 2803 2826 2813 2754 2682 2604 2586 2582 2639 2694 2788 ASK 2804 2814 2848 2882 2894 2878 2886 TT 3052 3089 3090 3078 3066 3065 3107 S222 3285 3308 3266 3253 3333 3398 3444 S991 3583 3606 3745 3759 3747 3737 3764 3870 3967 4099 4162 4282 4385 4554 4672 4854 5014 5103 52416 5435 5533 5644 727 5818 5886 6009 6285 6536 6753 674 6249 5345 4449 3665 1800 7044 85 76 7 3157 3016 4366 2635 4736 2859 2873 2824 2782 2787 2812 2794 26840 2848 2864 2844 2813 2811 2801 2734 2708 2700 2689 2665 2640 2657 2670 2686 2700 2724 2760 2782 2756 2747 2731 2705 2685 2680 2644 2666 2646 2852 2653 2659 2735 are 2705 2732 27935 2762 2769 2739 2753 2776 2835 2872 2899 2971 2964 3027 3024 2992 2982 3000 2994 2989 2988 3042 3098 3118 3114 3154 3211 3239 3300 3375 3424 3418 3354 3379 3322 3308 3360 3400 3352 3508 3633 HA 3851 3946 4106 4237 4361 4460 4528 4570 4598 4623 4689 4726 4651 4461 4162 3770 3359 1801 Theses 76 Syf/ 19 3770 3172 5564 2528 7004 2804 2766 2734 2704 2693 2749 2696 2662 2633 2548 2578 2643 2699 2696 2703 2708 2704 2689 2686 2695 2713 2740 2763 2788 2811 2821 2812 2834 2862 2850 2866 2895 2933 2780 3022 3026 3014 3004 2993 3021 3044 3041 3085 3088 3097 $125 3108 3143 3166 3169 3197 3258 3311 3350 3393 3458 3490 3518 3624 3643 3688 3747 3778 3834 3877 3926 4005 4047 4134 4219 4263 4330 4342 4380 4482 4533 4599 4704 4855 5018 5209 $441 5704 5930 6093 6260 6446 6643 6746 6877 5994 6965 6798 6440 59747 5277 4689 4167 3713 3328 MEAN : .94 77 48 , 35 34 52 625 62 S,D.:! ,20 46 07 .03 10% 06 04 40 MSENS :2 ,23 4.26 1.83 4.44 4.03 4.24 à :07 {37 *.*K as in Appendix 1 16,D, = Standard deviation 2MSENS = Mean sensitivity APPENDIX 3: STATISTICS FOR INTRA-RING WIDTH AND DENSITY VARIABLES FOR EACH CORE FROM NINE CRI LAKE WHITE SPRUCE TREES. Rw EU * L.W* RD* Bot? slp". oMnD™ * Mxp Tree 1, Core a YEARS: 4732-4975 MEAN: . 58 . 47 redial oes Tod »47 nies 57 S.D.:! 29 , 24 08 “0S 02 143 02 .40 MSENS :* iL? .26 57 09 06 136 5 (072 49 Tree 1, Core b YEARS: 1760-1977 MEAN: .55 .4i , 1. 4 Soe: . 34 Or : 30 .67 Se). 23 . 18 alle: .03 mee a htt) sie 1 MSENS : 19 , 25 38 . 08 .04 a? . 07 ,18 Tree 2, Core a YEARS; 4720-1977 MEAN: 46 a al, Ss 138 "Lie : 50 128 O0 S,D.: AL 49 44 , 05 03 SA 03 pee Gul MSENS: ned at 136 UD NEA . 18 . 08 ali Tree 2, Core b YEARS: 4720-1977 MEAN : . 55 , 45 410 . 34 .3 , 4S 126 57 SD: eZ 120 07 .03 .02 LS .03 .40 MSENS : . 20 .28 . 64 . 08 .06 43 YA aa Tree 3, Core a YEARS: 4790-1977 MEAN: AT. 67 . 40 Pet .27 . 44 es , 55 SD: 45 . 39 . DE 02 .02 aes .02 . 08 MSENS : .20 veo ‘bi. Oe .0S 42 .06 146 Tree 3. Core b YEARS: 41800-1977 MEAN: 169 62 .07 . 30 .28 135 a4 , 50 SALES . 38 . 36 0% 02 .02 REA De , 08 SENS : 1e 22 UE . 08 , DS .78 06 LA X,XX as in Appendix f{ 5.0, = Standard deviation 2MSENS = Mean sensitivity 180 APPENDIX 3: (Cont'd) RU* EUX LW * RD* Ept* LD oMND** “MKD Tree 4, Core a VARS 77101977 MEAN: .70 , 54 46 36 ae . 48 .28 .63 SOIT a , 24 .14 . 04 10e 416 .02 LS SENS :? a ol eu mee moe . 06 , 4. .08 ,18 Tree 4, Core b YEARS: 1770-1977 MEAN: . 64 SE ohh . 34 54 Bie) 27 nie Sibi eae nee , 14 06 .03 3 Ee .03 5 ALS MSIENS: ei 24 78 . 08 , 05 &1 .0S 18 Tree 5, Core a YEARS: 1730-1949 MEAN: eval . 54 HS AA noe Te Fr ET , 65 Sie eed ed . 0: .03 02 sibel 04 eis MSENS : AE OMC . 44 . 08 . 05 124 .06 .18 Tree S, Core ht WEARS si EDS MEAN: .70 139 ne , 34 . 30 146 5 AS bi SD LE 7 06 .03 102 iF 0e Drrte) MSENS : AS 16 . 66 LD .05S 46 07 eas Tree 6, Core a, First part YEARS: 4700-1900 MEAN: 90 6e eae .40 .34 ee ,29 .70 S.D.:! 22 ,20 AS 04 02 09 .02 ii MSENS : .49 2S .40 .09 0S AS .06 47 Tree 6, Core a, Second part YEARS: 4904-41977 MEAN: . 69 .58 14. . 33 30 .45 QE) 59 SD: 1 ao »06 .03 02 418 .02 . ii MSENS : 16 oui 68 .10 05 49 . 07 ed Tree 6, Core b YEARS: 1700-1977 MEAN: «73 .56 «17 38 133 .49 129 62 S.D.: 129 122 ‘ii .03 02 Date .02 .40 MSENS : .18 125 49 .09 :0S eb 06 .18 Tree 7, Core a YEARS? S2900-—-1977 MEAN: 14,12 4.00 13 32 7) , 44 625 06 SG, D, : . 30 se? . 08 .05 .02 43 02 ‘1e MSENS : .20 22 82 0 . 05 55 06 124 181 APPENDIX 3: (Cont'd) RW* Ew * Lw* R D** ED* LD HN MX * Tree 7, Core b YEARS: 1850-1977 MEAN: 120 x Lg Url Aa ME ea? 24S »24 59 Sipe: .38 .36 40 .03 02 17 .02 .12 MSENS ? Sir .20 .80 9 WU? 05 . 49 . 07 NUE) Tree 8, Core a YEARS: 4910-1977 MEAN: . 58 192 06 eae vee me 12e . 48 S.D.: ake! Le .05 mcs 02 ‘De .0i AL) MSENS : 15 .16 4.00 20 .06 . 38 .06 23 Tree 8, Core b YEARS: 1760-1977 MEAN : Br dat ‘bi . 0 ‘Ge ep .4i . 24 , 56 Sidi: ee 19 10 .04 .02 20 .02 US MSENS : 16 wy 88 ns ,05 107 .05 PL Tree 9, Core a YEARS: 4790-4977 MEAN : DE . 60 eue ‘33 . 30 . 44 . 24 .58 Se Dat «34 .28 .09 .03 (EE ni 0e peint MSENS : 17 125 US .08 .05 AE "07 a Tree 9, Core b YEARS: 4780-1903 MEAN: 193 .83 ran .28 . 26 , 30 el 49 Shit . 48 a? 44 .04 0: 3 .02 Ut MSENS : ree , 24 4.02 .08 .05S . 88 , 06 red 182 | APPENDIX 4: CRI LAKE COMPOUND-INCREMENT-FUNCTION (CIF) RING-WIDTH INDICES, "B & C" CHRONOLOGY (FIGURES 18, 19). : YEAR 0 1 2 3 4 5 6 7 8 9 | a SSS 1.00 1.05 1.16 1.06 1.05 0.97 1.06 1.11 1.01 1.28 0.86 1.28 157 1.35 1.06 1.31 0.87 0.95 0.88 1.09 1.40 1.34 0.94 1.21 1,22 1.47 1.37 1.65 1.55 1.41 1.10 1.10 1.10 1.38 1.26 0.88 1.25 1.09 0.91 1.13 1.02 1.08 1.10 1.18 1.00 115 1.16 1.06 0.90 1.02 1.04 1.41 1.24 0.89 0.96 0.86 0.89 1.17 1.06 + 0.92 0.97 0.79 0.90 1-22 0.75 0.92 1.06 0.92 1.31 1.29 1.00 0.88 0.74 0.87 0.78 0.89 0.82 0.90 0.82 1.02 1. 12 1.15 1.28 0.87 0.87 1.01 0.92 1.25 1.45 1.36 1.37 1:37 1.25 1.24 1.29 1.63 1.17 1.49 0.97 1.52 0.96 1.03 1.24 1.25 1.50 1.55 1.55 1.66 1.37 1.28 1.21 1.22 1.29 1.22 1.60 1.41 0.70 0.74 0.62 0.91 0.89 0.90 0.83 0.67 0.45 0.49 0.76 0.56 0.67 0.74 0.82 0.66 0.90 0.94 0.93 0272 0.85 0.76 0.82 0.98 0.85 0.95 0.82 0.50 0.65 0.72 0.87 0.85 0.79 0.98 0.97 1.05 1.04 1.45 1.20 1.06 1.21 0.93 1.02 0.87 0.83 0.97 0.98 12 1.15 1.14 0.88 1.00 1.12 0.74 0.94 0.79 0.95 0.91 1.19 1.01 1.20 1.14 1.08 0.96 1,12 1.14 1.24 1.08 1.10 1.18 1.18 1.34 1,21 1.00 1.10 1.10 1.06 1.17 1.14 1.31 1.09 1.01 0.92 0.91 0.96 1:13 0.97 3.11 1.11 1.12 1.20 1.25 1447 1.06 1.17 1.5 0.82 1.02 1.04 0.91 1.02 1.26 1.03 1.17 0.86 1.13 1.01 1.46 1.35 1.11 0.87 0.95 1.19 0.79 1.16 1.08 1.03 0.94 0.82 0.89 0.73 0.89 0.82 0.88 0.83 1.07 1.06 0.96 1.03 0.86 0.97 1.09 1.27 0.79 0.63 0.99 0.94 0.97 1.10 1.04 1.12 1.27 1.30 1.43 1.28 1.15 1,17 1.45 1.01 0.82 1.30 1.28 1.20 1.18 1.21 1.07 1.26 1.92 1.19 0.96 0.88 0.68 - = 183 APPENDIX 5: CRI LAKE MAXIMUM-DENSITY COMPOUND-INCREMENT-FUNCTION (CIF) INDICES, "B & C" CHRONOLOGY (FIGURES 18, 19). YEAR 0 1 2 3 4 5 6 a 8 9 1700 6.90 1.18 1.03 1.16 1.07 1.08 1.14 1.24 0.95 0.85 1710 1.22 1.12 1.19 1.10 0.98 1.24 0.80 1.05 0.87 0.94 1720 0.88 0.99 1.15 1.11 1.05 1.04 0.95 1.03 1.11 1.25 1730 0.91 0.98 1.09 1.23 1.06 1.19 1.17 1.04 1.14 1.05 1740 0.81 1.00 0.98 1.07 1.08 1.03 0.91 1.08 1.15 1.13 1750 0.89 1.23 1.03 0.90 1.11 0.79 1.03 0.69 0.83 0.86 1760 0.90 0.99 0.97 1.28 115 0.80 1115 0.91 1.11 1.14 1770 1.10 0.85 0.88 0.84 1.01 0.90 1.09 0.96 1.11 1.05 1780 0.99 1.24 1.25 1.10 0.83 1.04 0.96 0.88 1.15 1.22 1790 0.95 1.12 0.95 1.19 1.09 1.27 1.05 1.27 1.25 1.20 1800 0.83 1.12 1.05 1.08 1.31 1.26 1.01 0.91 1.06 1.04 1810 1.31 1.07 1.01 1.07 0.79 0.92 0.68 0.63 1.09 0.77 1820 1.08 1.04 0.86 0.91 0.81 0.96 1.10 0.75 1.07 0.96 1830 0.89 1.02 0.82 0.86 1.00 0.79 0.85 1.01 0.79 1.03 1840 1213 1.03 0.77 0.79 1.03 0.78 1.18 0.96 1.03 1.10 1850 1.08 1.03 1.02 0.71 0.92 0.88 1.18 0.96 1.16 0.83 1860 1.03 0.90 0.77 0.84 0.98 0.85 1.09 1.07 0.95 0.87 1870 1.15 0.77 1.16 1.14 1.04 1.28 1.27 1.25 1.28 1.21 1880 0.99 1.26 1.08 0.93 0.77 0.86 1.15 1.06 0.79 1.13 1890 0.93 0.84 1.14 1.19 0.93 1.20 1.12 1.11 1.02 1.16 1900 1.05 1.19 0.97 1.04 0.93 0.88 1.02 1.03 1.05 0.95 1910 1.04 1.28 0.72 0.99 0.90 1.02 1.16 0.95 0.77 1.29 1920 1.12 1.05 1.14 0.79 1.17 0.82 0.95 1.08 1.15 0.87 1930 1.18 0.99 1.08 0.97 0.72 1.10 0.89 1.14 1.14 1.10 1940 1.13 0.93 1.10 0.97 1.08 0.93 0.85 0.94 1.18 1.08 1950 0.96 1.05 1.05 1.01 1.37 1.27 0.67 0.82 0.83 1.17 1960 1.08 1.10 1.06 0.76 1.08 0.78 1.19 1.06 1.09 0.69 1970 1.19 0.96 0.94 1.27 1.14 1.11 0.94 1.06 = = 184 "B" CHRONOLOGY (FIGURES 18, 20). APPENDIX 6: CRI LAKE 13-YEAR WEIGHTED-RUNNING MEAN OF COMPOUND-INCREMENT-FUNCTION (CIF) RING-WIDTH INDICES, YEAR 0 1 2 3 4 5 6 7 8 9 1700 1.06 1.07 1.07 1.06 1.05 1.05 1.06 1.07 1.09 Wis 12 1710 1.17 1.23 1.27 1.26 1.20 1.12 1.05 1.03 1.05 1.11 1720 1.16 fe 19 1.20 1.23 1.29 1.36 1.43 1.45 1.42 1.35 1730 1.27 1.21 lle £9 1.18 1.16 1.13 1.10 1.08 1.06 1.06 1740 1.07 1.08 1.09 1.10 1.10 1.10 1.08 1.06 1.05 7 1.07 1750 1.11 1.13 1.10 1.05 0.99 0.97 0.98 0.99 0.99 0.97 1760 0.95 0.94 0.94 0.95 0.96 0.97 1.01 1.07 1.10 1.09 1770 1.02 0.94 0.88 0.84 0.84 0.85 0.86 0.89 0.94 1.00 1780 1.05 1.08 1.06 1.03 1.00 1.03 1.10 te 12 1.28 1.33 1790 1.34 1.33 1.32 1.32 1.34 1.35 1.33 1.30 1.25 1.21 1800 1.18 1.18 1.23 1.32 1.41 1.48 1.50 1.47 1.41 1.34 1810 1.29 1.28 1.29 1.29 1.25 1.14 1.00 0.88 0.82 0.82 1820 0.83 0.81 0.75 0.68 0.63 0.61 0.62 0.65 0.68 0.72 1830 0.76 0.80 0.83 0.85 0.85 0.84 0.83 0.83 0.85 0.87 1840 0.86 0.82 0.77 0.73 0.72 0.74 0.79 0.84 0.88 0.93 1850 0.98 1.02 1.07 1.10 1.11 1.10 1.07 1.02 0.98 0.95 1860 0.95 0.98 1.03 1.07 1.08 1.07 1.03 0.99 0.95 0.92 1870 0.90 0.91 0.94 0.99 1.04 1.08 1.10 1.10 1.09 1.09 1880 1.10 1.12 1.14 1.14 1.15 Ter 17 1.19 1.19 1.17 1.13 1890 1.11 1.10 1.12 1.14 1.15 1.14 1.09 1.04 1.00 0.98 1900 1.00 1.02 1.05 1.08 1.11 1.14 1.16 1.17 1.16 1.13 1910 1.10 1.06 1.02 1.00 1.00 1.02 1.05 1.07 1.08 1.07 1920 1.08 1.11 1.16 1.19 1.18 1.13 1.06 1.02 1.01 1.02 1930 1.03 1.02 0.99 0.94 0.89 0.86 0.84 0.84 0.86 0.89 1940 0.92 0.96 0.99 0.99 0.98 0.99 1.00 1.01 0.98 0.94 1950 0.90 0.91 0.94 0.99 1.05 1.10 1.16 1.22 1.27 1.28 1960 1.27 1.25 1.21 1.18 1.14 1.13 1.15 1.18 1.20 1.19 1970 1.18 1.17 1.15 1.12 1.06 0.97 0.88 0.78 = - 185 APPENDIX 7: CRI LAKE 13-YEAR WEIGATED-RUNNING MEAN OF COMPOUND-INCREMENT-FUNCTION (CIF) MAXIMUM-DENSITY INDICES, "B" CHRONOLOGY (FIGURES 18, 20). YEAR 0 1 2 3 4 5 6 7 8 9 1700 1.02 1.05 1.08 1.09 1.10 1.11 1.10 1.08 1.07 1.06 1710 1.08 1.10 1.11 1.10 1.07 1.03 1.00 0.96 0.95 0.95 1720 0.97 1.01 1.04 1.06 1.05 1.04 1.04 1.05 1.07 1.07 1730 1.06 1.07 1.09 dti 1.13 1.13 1:12 1.09 1.06 1.02 1740 0.99 0.99 1.00 1.02 1.03 1.03 1.04 1.05 1.07 1.07 1750 1.07 1.05 1.03 1.00 0.96 0.92 0.89 0.86 0.86 0.88 1760 0.92 0.98 1.03 1.06 1.06 1.04 1.03 1.04 1.05 1.04 1770 1.01 0.96 0.93 0.92 0.94 0.97 1.00 1.03 1.05 1.07 1780 1.10 1.12 1.11 1.07 1.02 0.99 0.99 1.02 1.05 1.07 1790 1.07 1.07 1.08 1.11 1.13 1.16 1.17 1.18 1.16 1.12 1800 1.09 1.08 1.10 1.13 1.14 1.13 1.09 1.07 1.07 1.08 1810 1.10 1.08 1.03 0.97 0.90 0.85 0.82 0.83 0.87 0.91 1820 0.94 0.95 0.93 0.92 0.92 0.93 0.94 0.95 0.95 0.95 1830 0.94 0.92 0.91 0.90 0.89 0.89 0.90 0.91 0.94 0.97 1840 0.97 0.95 0.92 0.91 0.92 0.95 0.99 1.02 1.04 1.04 1850 1.03 1.00 0.96 0.93 0.93 0.96 1.00 1.01 1.00 0.97 1860 0.94 0.90 0.89 0.89 0.92 0.95 0.98 0.99 0.99 0.98 1870 0.99 1.02 1.05 1.10 1.15 1.19 1.22 1.23 1.21 1.18 1880 1.14 1.09 1.04 0.98 0.95 0.95 0.97 0.98 0.98 0.98 1890 0.98 1.01 1.04 1.07 1.09 1.10 1.10 1.10 1.10 1.09 1900 1.09 1.07 1.04 1.01 0.98 0.98 0.99 1.00 1.02 1.03 1910 1.03 1.01 0.98 0.97 0.97 0.99 1.00 1.01 1.03 1.06 1920 1.07 1.06 1.04 1.01 0.99 0.98 1.00 1.02 1.03 1.04 1930 1.04 1.02 1.00 0.97 0.96 0.97 1.01 1.05 1.08 1.08 1940 1.07 1.05 1.03 1.01 0.99 0.98 0.98 1.00 1.02 1.04 1950 1.05 1.06 1.08 1.10 1.09 1.05 0.98 0.94 0.96 1.00 1960 1.03 1.03 1.00 0.98 0.97 0.99 1.01 1.01 1.00 1.00 1970 1.00 1.03 1.06 1.09 1.10 1.08 1.07 1.06 = = 186 APPENDIX 8: CRI LAKE 13-YEAR DIGITAL-FILTER RING-WIDTH INDICES, "C" CHRONOLOGY (FIGURES 18, 21). YEAR 0 1 2 3 4 5 6 7 8 9 1700 0.94 0.98 1.09 1.00 1.00 0.92 1.00 1.04 0.92 1.14 1710 0.73 1.04 1.24 1.07 0.88 1217 0.82 0.93 0.84 0.98 1720 1.20 1.12 0.78 0.98 0.95 1.08 0.96 1.14 1.09 1.05 1730 0.87 0.91 0.92 1" 1.08 0.78 1.14 1.01 0.86 1.07 1740 0.96 1.00 1.01 1.07 0.91 1.05 1.07 1.00 0.86 0.95 1750 0.94 1.25 1.12 0.85 0.97 0.89 0.91 1.18 1.07 > 0.95 1760 1.02 0.84 0.96 1.29 0.79 0.94 1.04 0.86 1.19 1.19 1770 0.98 0.93 0.84 1.03 0.93 1.05 0.95 1.01 0.87 1.02 1780 1.06 1.07 1.20 0.85 0.87 0.99 0.84 1.05 1.14 1.02 1790 1.02 1.03 0.95 0.94 0.96 121 0.88 1.15 0.78 1.26 1800 0.82 0.87 1.01 0.95 1.07 1.05 1.03 1.13 0.97 0.95 1810 0.94 0.96 1.00 0.94 1.28 1.23 0.70 0.84 0.76 1.11 1820 1.08 nt 1.10 0.98 0.72 0.81 1.23 0.87 0.98 1.03 1830 1.08 0.83 1.08 1.10 1.09 0.86 1.03 0.91 0.97 LL 1840 0.99 1.15 1.06 0.69 0.91 0.97 1.10 1.02 0.90 1.05 1850 0.99 1.02 0.98 1.03 1.08 0.96 1.13 0.91 1.04 0.92 1860 0.88 0.99 0.95 1.15 1.06 1.07 0.86 1.01 1.17 0.80 1870 1.04 0.87 1.01 0.92 1.14 0.93 1.09 1.04 0.99 0.88 1880 1.02 1.02 1.09 0.94 0.95 1.01 0.99 1.13 1.04 0.88 1890 0.99 1.00 0.95 1.03 0.99 1.15 1.00 0.97 0.92 0.92 1900 0.96 1.10 0.92 1.03 1.00 0.98 1.03 1.07 1.01 0.94 1910 1.07 1.09 0.80 1.02 1.04 0.89 0.97 1.17 0.95 1.09 1920 0.80 1.02 0.87 1:23 1.14 0.99 0.82 0.93 1.17 ag 1930 13 1.06 1.04 1.00 0.92 1.04 0.87 1.06 0.96 0.99 1940 0.90 41 1.08 0.97 1.05 0.87 0.97 1.08 1.29 0.84 1950 0.70 1.09 1.00 0.98 1.05 0.95 0.96 1.04 1.03 1.11 1960 1.01 0.92 0.96 1.23 0.88 0.72 1.13 1.08 1.00 0.99 1970 1.03 0.92 1.09 1.00 1:12 0.99 1.00 0.87 - - 187 APPENDIX 9: CRI LAKE 13-YEAR DIGITAL-FILTER MAXIMUM-DENSITY INDICES, "C" CHRONOLOGY (FIGURES 18, 21). YEAR 0 1 2 3 4 5 6 7 8 1700 0.88 1.12 0.96 1.06 0.97 0.98 1.04 ee 0.89 1710 1.13 1.01 1.07 1.00 0.92 1.20 0.80 1.09 0.92 1720 0.91 0.98 1.10 1.05 1.00 1.00 0.91 0.98 1.04 1730 0.86 0.92 1.00 1.11 0.94 1.05 1.04 0.95 1.08 1740 0.82 1.01 0.98 1.05 1.05 1.00 0.88 1.03 1.08 1750 0.84 1.17 1.00 0.90 1.15 0.86 1.16 0.80 0.97 1760 0.97 1.01 0.94 1.21 1.09 0.77 owl 0.88 1.06 1770 1.09 0.88 0.94 0.91 1.07 0.93 1.09 0.94 1.06 1780 0.90 1.11 1.13 1.03 0.81 1.05 0.97 0.86 1.10 1790 0.89 1.04 0.88 1.08 0.96 1.10 0.89 1.08 1.08 1800 0.76 1.04 0.96 0.96 vows 1.12 0.92 0.85 0.99 1810 1.20 0.99 0.98 1.10 0.87 1.09 0.83 0.76 1.26 1820 1.15 1.10 0.92 0.99 0.88 1.03 1.17 0.79 1.13 1830 0.95 1.11 0.90 0.96 1.12 0.89 0.95 1.10 0.84 1840 1.16 1.08 0.84 0.87 1.12 0.82 1.19 0.94 0.99 1850 1.05 1.03 1.07 0.77 0.99 0.92 1.19 0.95 1.16 1860 1.10 1.00 0.87 0.94 1.07 0.89 1.11 1.08 0.96 1870 1.16 0.76 1.10 1.03 0.90 1.07 1.04 1.02 1.05 1880 0.87 1.15 1.04 0.95 0.81 0.90 1.19 1.08 0.80 1890 0.94 0.83 1.10 1.11 0.85 1.09 1.02 1.01 0.93 1900 0.97 1.11 0.93 1.03 0.95 0.90 1.03 1.03 1.03 1910 1.01 1.27 0.73 1.02 0.92 1.03 1.16 0.94 0.75 1920 1.04 0.99 1.10 0.79 1.19 0.83 0.95 1.06 oaks 1930 1.14 0.97 1.08 1.00 0.75 15 Je} 0.88 1.08 1.06 1940 1.06 0.89 1.07 0.96 1.09 0.95 0.87 0.94 1.15 1950 0.92 0.99 0.97 0.92 1.25 1.21 0.68 0.87 0.87 1960 1.05 1.07 1.06 0.78 1.11 0.79 1.18 1.05 1.09 1970 119 0.94 0.89 1.17 1.04 1.03 0.88 1.00 = = 188 IMPACT OF CLIMATIC VARIATION ON BOREAL FOREST BIOMASS PRODUCTION John M. Powell* INTRODUCTION Climate has a direct and pronounced effect on forest vegetation. However, there is a need to establish the impact of climate, especially climatic fluctuations, on forest biomass productivity. Little is known about the effect of climatic factors on biomass and how climatic fluctuations effect forest growth and yield. Renewed interest in the use of forest products and wastes for energy supplements as liquid fuels, and in using certain forest products for food, especially animal fodders and supplements, demonstrate there is a need to develop long-term inventories of biomass in forested areas and to establish projected annual yields. Such inventories and projections must take climatic factors into account. The rate of biomass renewal under a wide range of environmental, site and growing stock conditions is unknown. Present day forests and their associated vegetation did not evolve under constant or uniform climatic conditions. Different regions of our forests are subjected to widely contrasting temperature and precipitation regimes. Weather patterns are highly variable from year to year. Today's forests reflect past climates. Their origins in time and space, growth rates, species composition, longevity and total biomass are largely dictated by past and present climates. Predictions relating to future forest vegetation or biomass must include a consideration of climatic change. Some climatologists believe that North American climate is unlikely to remain as amenable in the near future as during several recent decades - a period of relatively stable climate (Hare 1979; Science Council of Canada 1976). Some predict a significant cooling, others a warming, but most indicate there will be a return to increased climatic variability with subsequent impact on energy, food and other resources (National Academy of Sciences 1976; Norwine 1977; Science Council of Canada 1976). Se. El * Northern Forest Research Centre, Canadian Forestry Service, Enviroment Canada, Edmonton, Alberta, T6H 3S5 189 Past decisions concerning the use of Boreal Forest and adjacent areas have involved little thought as to the impact of changing climate on long-term productivity of the resource. Continuing disregard to climatic variations could, therefore, have severe social and economic implications. Clearly, we need more information on climatic fluctuations that have occurred in the Boreal Forest zone. However, there are few long-term weather stations in this zone, and none with a record of 100 years. Some proxy data* indicate that important climatic fluctuations occurred in the past, and that the position of the Boreal Forest zone and northern tree line has shifted also (Kay 1978, 1979; Larsen 1965, 1974; Nichols 1967, 1969, 1975, 1976; Ritchie 1976; Ritchie and Hare 1971; Ritchie and Yarranton 1978; Sorenson 1977; Sorenson and Knox 1974). The Boreal Forest presently occupies a zone with a summer temperature range of only 2°C. Any long-term cooling, even of 0.5°C, would adversely affect this zone and its biomass productivity. Therefore, information is needed in two specific areas: (1) impact of climate on forest biomass production; and (2) past climates in the Boreal Forest region. OBJECTIVES AND APPROACHES Objectives The two overall objectives of the study are to determine: (1) extent and degree of past short- and long-term climatic variation in selected regions of the Boreal Forest to assess impact of climatic change upon tree and forest growth; (2) quantitative and qualitative relationships between key climatic parameters/climatic variations, and measures of forest biomass growth and production in selected regions of the Boreal Forest. * Data drawn from biological, chemical and physical characteristics of the earth's environment that respond to climate, and that are useful in the reconstruction of past climates. 190 Approaches All long-term instrumental climatic records available for the central Boreal Forest zone (Yukon to northwestern Ontario) will be analyzed to establish regional climatic (particularly temperature and precipitation) trends and fluctuations. In addition, proxy data relating to the Boreal Forest zone will be used (e.g. data from dendrochronology, paleobotany, lake sediments including foraminifera, speleothems, isotopes, archaeology and historical documents). This will help to establish climatic variation over a longer period. To assist with this review, a two-year contract has been awarded to Western Ecological Services Ltd. to carry out an extensive literature survey of growth, yield and biomass of different tree species and other vegetation in the Boreal Forest zone, with particular emphasis on North America. The effect of temperature, precipitation, etc. on growth and yield of biomass will be noted. The survey will include references to dendrochronology (dendroclimatology) and other proxy data which may be useful in establishing relationships between biomass (phytomass) productivity and climatic variations. A preliminary evaluation of qualitative and quantitative relationships between key climatic parameters and measures of biomass productivity in the Boreal Forest will also be included. Tree-ring analysis can provide answers in both areas indicated at the end of the Introduction. Tree-ring chronologies have long been known to be useful in gathering information on past climates and dateable phenomena, providing normal tree growth trends are considered. This type of analysis can also help identify various climatic factors that have played a major role in forest growth and biomass production. Ring-width and density profiles can also provide year-to-year biomass data. Dendroclimatological techniques have been successfully applied in western Canada by Parker and his colleagues (Parker 19/76; Parker and Jozsa 1973, 1977; Parker and Kennedy 1973; Parker, Schoorlemmer and Carver 1973; Parker, Bruce and Jozsa 1977). Parker has shown that on an altitudinal transect in the Columbia River valley of southeastern British Columbia, temperature/tree-ring correlations were more significant at higher elevations, and precipitation/tree-ring correlations at lower elevations (Western Forest Products Laboratory 1979, pp. 23-24). A similar relationship may occur on a latitudinal transect of the Boreal Forest zone. If this can be shown, it is relevant to any dendroclimatological studies undertaken in the Interior Plains and elsewhere in North America. 191 A two-year contract has been awarded to Parker and his group at Forintek Canada Corporation to undertake a pilot study of tree-ring chronologies and the impact of climatic change on Boreal Forest biomass productivity. Field work will include the selection of five or more sampling sites along two north-south transects, from near the northern tree-line through the Boreal Forest zone to the Aspen Parkland, using white spruce (Picea glauca) as the sample species. One transect would be in the Alberta/Northwest Territories area near longitude 115°W, and the other near the Saskatchewan/Manitoba boundary. In September 1979, 300-year-old tree samples were obtained from the Swan Hills area of Alberta; in March 1980, 300-year-old samples were obtained from near Fort Resolution in the Northwest Territories. In August 1980, 200 + -year-old samples were collected from west of Sundre and north of Hinton, Alberta, and in September 1980 from west of Fort Vermilion, Alberta and from just north of Lake Claire in Wood Buffalo National Park. In September, 200 + year-old samples were also collected from the Cameron Hills southwest of Hay River, and at Prelude Lake east of Yellowknife in the Northwest Territories. Sampling along the second transect will be undertaken in 1981. Disks are being obtained from 10 to 15 trees at each site on the transect and at various heights on the trees. Two average radii from each disk will be prepared for x-ray densitometry analysis to derive ring-density chronologies and biomass data. The data will be placed on tape using various programs, so that all data can be subsequently analyzed for various climatic factors from which growth trends have been removed. This will result in a record of processed tree-ring widths and densities suitable for climatic and biomass analyses. Presumably these north-south transects through the Boreal Forest will tie in with studies by G.C. Jacoby* and others near the northern tree-line, providing a network which could lead to more comprehensive climatic analyses of large areas of central Canada. This, in turn, could add to the value of isolated, local reconstructions of past climate based on various proxy sources. * See papers by Jacoby and Ulan, and by Parker et al. in this publication. 192 REFERENCES Hare, F.K. 1979. Climatic variation and variability: empirical evidence from meteorological and other sources. In: Proceedings of the World Climate Conference, a Conference of Experts on Climate and Mankind. World Meteorol. Org., Geneva, WMO-No. 537. pp. 51-87. Kay, P.A. 1978. Dendroecology in Canada's forest-tundra transition zone. Arctic and Alpine Res. 10:133-138. . 1979. Miltivariate statistical estimates of Holocene vegetation and climate change, forest-tundra transition zone, N.W.T., Canada. Quaternary Res. 11:125-140. Larsen, J.A. 1965. The vegetation of the Ennadai Lake area, N.W.T.: studies in Subarctic and Arctic bioclimatology. Ecol. Monogr. 35:37-59. - 1974. Ecology of the northern continental forest border. In: Arctic and Alpine Environments. Edited by: J.D. Ives and R.G. Barry. Methuen, London. pp. 343-369. National Academy of Sciences. 1976. Climate and food; climatic fluctuation and U.S. agricultural production. Washington, D.C. 212 pp. Nichols, H. 1967. The post-glacial history of vegetation and climate at Ennadai Lake, Keewatin, and Lynn Lake, Manitoba (Canada). Eiszeitalter und Gegenwart 18:176-197. - 1969. The late Quaternary history of vegetation and climate at Porcupine Mountain and Clearwater Bog, Manitoba. Arctic and Alpine Res. 1:155-167. - 1975. Palynological and paleoclimatic study of the late Quaternary displacement of the Boreal Forest-tundra ecotone in Keewatin and Mackenzie, N.W.T., Canada. Univ. of Colorado, Inst. Arctic and Alpine Res., Occasional Paper No. 15:1-87. . 1976. Historical aspects of the northern Canadian treeline. Arctic 29:38-47. Norwine, J. 1977. A question of climate. Environment 19(8):7-13, 25-27. Parker, M.L. 1976. Improving tree-ring dating in northern Canada by x-ray densitometry. Syesis No. 9:163-172. Parker, M.L., R.D. Bruce, and L.A. Jozsa. 1977. Calibration, data acquisition and processing procedures used with an online tree-ring scanning densitometer. IUFRO Group P4.01.05, Instruments Meeting, Corvallis, Oregon, Sept. 8-9, 1977. 20 pp. Parker, M.L., and L.A. Jozsa. 1973. X-ray scanning machine for tree-ring width and density analysis. Wood and Fiber 5:192-197. . 1977. What tree rings tell us. Fisheries and Env. Can., Can. Forestry Serv. Fact Sheet. 4 pp. Parker, M.L., and R.W. Kennedy. 1973. The status of radiation densitometry for measurement of wood specific gravity. Proc. IUFRO, Div. 5 Meetings, Cape Town and Pretoria, South Atrica. Li pp. Parker, M.L., J. Schoorlemmer, and L.J. Carver. 1973. A computerized scanning densitometer for automatic recording of tree-ring width and density data from x-ray negatives. Wood and Fiber 5:237-248. 193 Ritchie, J.C. 1976. The late-Quaternary vegetational history of the Western Interior of Canada. Can. J. Bot. 54: 1793-1818. Ritchie, J.C., and F.K. Hare. 19/71. Late-Quaternary vegetation and climate near the Arctic tree line of northwestern North America. Quaternary Res. 1: 331-342. Ritchie, J.C., and G.A. Yarranton. 1978. The late-Quaternary history of the Boreal Forest of central Canada, based on standard pollen stratigraphy and principal components analysis. Jie HCO. Go: 199212 Science Council of Canada. 1976. Living with climatic change. Edited by: K. Beltzner. Proc. Toronto Conf. Workshop, 1975. Ottawa. 105 pp. Sorenson, C.J. 1977. Reconstructed Holocene bioclimates. Ann. Assoc. Amer. Geographers Oise (214-222. Sorenson, C.J., and J.C. Knox. 19/74. Paleosols and paleoclimates related to Late Holocene forest/tundra border migrations: Mackenzie and Keewatin, N.W.T., Canada. In: International Conference on Prehistory and Paleoecology of the Western North American Arctic and Subarctic. Edited by: S. Raymond and P. Schledermann. Univ. of Calgary, Archaeol. Assoc. pp. 187-203. Western Forest Products Laboratory. 1979. Western Forest Products Laboratory program review 1977 and 1978. Env. Can., Can. Forestry Serv. 42 pp. 194 STUDYING CLIMATIC CHANGE FROM CANADIAN HIGH ARCTIC ICE CORES R.M. Koerner* and D.A. Fisher* INTRODUCTION The accumulation regions of ice sheets and ice caps are repositories of past atmospheres. This is because as snow falls in the upper and colder regions of ice sheets it includes with it, as it accumulates, some of the air at the surface and also aerosols from the atmosphere. The snow grains themselves contain impurities from the air gathered either as a nucleus or by a sweeping action during their fall at the time of precipitation. The ratios of the various isotopes in the water molecule (specifically DH (deuterium, an isotope of hydrogen), 160, 180) depend on the temperature at the condensation level when the snow falls. Therefore, if we take a core section in an area where the ice flow is least complex so that the accumulated layers are not mixed up by folding or faulting we can study both the changing temperature of condensation and atmospheric content back in time. The time scale resolution depends on the accumulation rate of snow at the surface and also on the degree of summer melt. So an ideal drill site should be at the top of the flow-line with an accumulation of about 20 g cm2 yt (grams per square centimetre per year) or more and very little melt. Some High Arctic ice caps provide such conditions. A record of over a million years of accumulation is claimed for ice in Antarctica. The High Arctic Canadian ice caps span much less than this, about 100,000 or 200,000 years. Ice coring has a relatively short history and began in earnest when the Norwegian-British-Swedish Expedition of 1950 and 1951 took a 100-m core in Antarctica. Since then several cores have been taken from Antarctica and Greenland by different nations. The two longest cores are surface-to-bedrock cores in Antarctica (Byrd Station (2164m)) and Greenland (Camp Century (1387m)). The latter is probably the most studied of all cores. * Polar Continental Shelf Project, Energy Mines and Resources Canada, 880 Wellington St., Ottawa, Ontario, K1A OE4 195 112° Ward Hunt 1. Meighe Meighe Dri I site gaftin ad 0 100 200 eee eed Scale of Kilometers 72° 64° FIGURE 1: Location of drill sites (black dots) in High Arctic Canada and northwestern Greenland. Grey areas denote glacial ice. 196 The Polar Continental Shelf Project began its drilling program in 1965 with an 112-m thermally drilled surface-to-bedrock core (M65) on Meighen Ice Cap (Figure 1). The program continued with a 212-m core on Devon Island Ice Cap in 1971 (D71), and two surface-to-bedrock cores 299-m in length from the same ice cap in 1972 and 1973 (D72, D73). Presently we are engaged in drilling on the Agassiz Ice Cap on Northern Ellesmere Island where, to date, a 337-m core (Ag77) and a 137-m core (Ag79) have been drilled. Both extend from the surface to bedrock. All these cores come from, or close to, the top of the flow-line on each ice cap. METHODOLOGY Most groups studying ice cores in other countries do so on an interdepartmental basis with each agency studying only one aspect of core chronology. At the Polar Continental Shelf Project most analyses are conducted and interpreted within that institution. The most notable exception is that almost all our oxygen isotope analyses are done by the Geophysical Isotope Laboratory at the University of Copenhagen. In the past, the University of Calgary (R. Krouse) and the University of British Columbia (D. Russell) have done oxygen isotope analyses as well. Oxygen Isotopes The ratio 180/169 in water or ice is determined by fractionation due to the lower (by 10%) vapour pressure of the heavier isotope. The basic relationships are well described in Dansgaard et al. (1973). The varying ratio with depth in ice cores is interpreted in terms of changing temperature at the drill site back in time via an empirical relationship where a 1°C change in temperature gives a 0.62 ©°/00 4 180 difference (Dansgaard et al. 1971, Koerner and Russell 1979). 6 here is the fractional difference between the ratio in the sample to that in standard mean ocean water, in fact, the NBS* number 1 standard. The 5/t (where t is temperature) relationship must be viewed cautiously as 6 will also alter according to changes in the precipitation source area, the ice thickness, the flow-line and the ratio of winter-summer precipitation. * National Bureau of Standards. We)7) Snow Chemistry Impurities in the air are removed either by gravitational (dry) fall out, by formation as precipitation nuclei (rain out) or by the sweeping action of rain and snow as it falls (wash out). Dry fall out is the least important of these mechanisms. The impurity concentrations in the snow (which in polar snows are generally in the low parts per billion (ppb) level) have been considered by some (e.g. Boutron and Lorius 1979) to be similar to those in the air. Junge (1977), however, has questioned this. Over long periods of time (i.e. more than one year), it is probably safe for us to assume that changes in the concentration of impurities in the snow parallel changes in the concentration of impurities in the air flowing over the ice cap. In our work, concentrations of Na, Ca, K, Al, Si and Mg are measured on a Perkin-Elmer 603 atomic absorption spectrophotometer. As the concentrations are in the ppb range, special care must be taken in preparing the ice sample before its final melt. We also measure total ion concentrations on a conductivity meter, and the acidity of the samples on a pH meter. Dust particles, being insoluble in water, are measured on a Coulter counter which can count particles as small as 0.5 um. In general, dust in our cores is less than 10 ym diameter and follows a log distribution according to the differential equation: aN/d (log r) = cr~§ (1) Where B is about 3, c is a constant and N is the total concentration of aerosols of radius smaller than r. The concentration of dust in our cores is about 250 ppb by weight. Site Selection Before drilling, it is essential to determine the nature of the topography under the ice cap of interest. As a result of our work between 1970 and 1980 we have found that the Canadian ice caps on Devon, Ellesmere and Axel Heiberg Islands range in thickness from about 150 m to 1000 m. Bedrock highs are usually associated with thin ice and bedrock valleys with thicker ice. As a consequence, most of the ice caps are thin (less than 300 m) in their higher regions (top of the flow-line) as they invariably coincide with bedrock highs. This limits our work as it means our ice cores, which are preferably taken right at the top of the flow-line where the flow is dominantly vertically downwards, are 300 m or less in length. 198 While this spans a period of time exceeding 100,000 years, only the last 10,000 years has a reasonable time resolution. Beyond that, our time scale is to imprecise to make more than generalized comments. RESULTS Meighen Island W.S.B. Paterson began the Polar Continental Shelf Project drilling program with a 112-m core on the Meighen Ice Cap in 1965 (M65). This ice cap consists entirely of ice formed each spring by melt water, which develops at the surface of the snow pack, percolating down and refreezing on the ice surface under the snow pack. Our analysis of the texture and fabric of the ice showed the ice cap has been stagnant throughout its history and has never been much bigger than it is at present. The same is probably true of most small ice caps in the Queen Elizabeth Islands (Koerner 1968). The value of the M65 core in terms of climatic change studies is seriously limited because in most years part of the snow pack melts and, by run off, is lost to the ice cap. In very warm summers (about 1 in 10 at present), one or more years of accumulation are lost. The climatic record is therefore discontinuous. Nevertheless, our studies have enabled us to draw up an approximate climatic history of the ice cap (Table 1). Evidently the ice cap originated after the climatic optimum, and therefore is less than 5,000 years old. There is evidence of a 1,000-year period of a negative balance at the drill site ending about 1,000 B.P. Subsequently, the ice cap grew again and experienced the coldest period in its 4,000-year history a few hundred years ago. This history is similar to that of the Ward-Hunt Ice Shelf on the north coast of Ellesmere Island (Lyons and Mielke 1973). Paterson (1968), by analyzing the temperature profile of the ice cap, concluded that the mean annual surface temperature increased by about 3.5°C between 1880 and 1940 AD and has decreased by about 1.5°C since then. Alt (1979), in an intensive study of summer synoptic climate controls of the Meighen Ice Cap mass balance, has been able to relate the general circulation of the atmosphere to the health of the ice cap. Dominance of the winter pattern of a 500-mb Low in the Hudson Bay-Baffin Island region throughout the summer season is capable of maintaining Meighen Ice Cap at its present size. A shift of the 500-mb Low from the winter position directly to the EQS TABLE 1: CLIMATIC HISTORY OF THE MEIGHEN ICE CAP, NORTHWEST TERRITORIES. Depth Interval (m) 24-44 44-54 54 54-116 116-121 200 Estimated Time Interval (years B.P.) Present - 80 80 - 390 390 - 560 560 - 660 660 - 2500/2000 2500/2000-? ?-3000/4500 Climate and Balance at the Core-site Ablation surface formed during period of negative balance which removed a maximum of 13 m of ice at the core site. Positive balance. Relatively little summer melt, coldest period in ice cap's history, positive balance. Core-site on a well-drained slope, positive balance. Long period of negative balance which reduced the ice cap in size. Overall positive balance with several short periods of negative balance. Positive balance, with the ice cap covering most of Meighen Island. Beaufort Sea or adjoining Polar Ocean area is capable of increasing the size of the ice cap. On the other hand, a shift of the 500-mb vortex to the Asiatic side of the Polar Ocean before taking up position in the Beaufort Sea - Polar Ocean area produces negative mass-balance conditions. When the 500-mb Low remains on the Asiatic side of the Polar Ocean during most of the summer season, the slow accumulation of two decades of Polar Ocean years is destroyed. Devon Island The Devon Island Ice Cap was chosen as our first deep drill site because of the smooth bedrock profile underlying its summit. In addition, there is an invaluable 20-year record of mass balance measurements for the ice cap against which to evaluate some of our results (Table 2). TABLE 2: ANNUAL MASS BALANCE OF THE DEVON ISLAND ICE CAP, NORTHWEST TERRITORIES (1961-1978). YEAR (SA "62" "63 “64 765) "66" 67> "68" 69" 7970 271) 72 750 74) SN ST TT Annual Mass Balance -20 -36 +4 +13 +6 -13 -3 -28 +4 -7 +10 -9 -8 -7 +17 -10 +3 g cm-2y~1 Time Scales Variations of any parameter within a long ice core would be valueless in terms of climatic change without a time scale for that core. Time scales may be calculated on the basis of simple ice flow models. For example, the relationship Caen A (2) Where t= 1 y, H is the thickness in metres, À is the accumulation rate in metres of ice measured at the surface, and y is the distance in metres of the layer from the ice/bed interface. Time scales are dealt with in Paterson (1980) and Hammer et al. (1978). Equation (2) is discussed in the second of these papers. 201 However, while such a time scale is satisfactory for the initial cutting of the oxygen isotope analysis and for a preliminary discussion of results, a more refined needed for final evaluation. In the Devon core (Paterson et al. 1977), we combined independent methods to determine a time scale. (1) Vertical strain measurements (Paterson 1976) in the bore hole over a period of give the vertical ice velocity field (v) with depth. v can then be integrated z and used in the relationship: t = f 2y-laz (3) o The total vertical ice velocity on a steady state ice cap must be equal to the cores for one is four time (t) by depth accumulation rate, otherwise, the ice cap will thicken or thin. v decreases with depth to zero at the bed, because in ice caps considered here the basal ice is frozen to the bedrock and therefore is motionless. The method assumes a steady state over the time interval in question, which may not be valid. (2) À (annual layer thickness) was measured by recording seasonal variations of microparticle concentrations (and of non-diffusing elements such as K or Al) at various levels in the core. A polynomial fit of À to z was then used to refine the time scale based on method (1). (3) A further check was then made using variations in the concentration of 32si and 14c in the core. In this case a knowledge of the concentration of these radioactive isotopes at the surface allows a calculation of (t) in: Az = Age-qt (4) Where A is the 14C or 32si specific radioactivity of the snow at the surface (Aj) or at depth z, q is the radioactive constant of 32si or 4c, (4) We have further checked our time scale by statistical comparisons of our 6 profiles with that from Camp Century, Greenland, where the time scale is considered correct within + 5% (Hammer et al. 1978). This final check indicated that our time scale (Paterson et al. 1977) was essentially correct. Therefore, our time scale is correct within + 5% from 0-5,000 B.P., but becomes increasingly unreliable beyond that. The Temperature Record of the Devon Island Ice Cap (D72 and D73) Cores The last 130,000 years 5 record is shown in Figure 2. It includes the cold period lasting from 10,000 to 60,000 B.P. that is considered to represent either the whole, or the major part of, the last (Wisconsin) glaciation. Sea-level records indicate lower sea levels 202 FIGURE 2: NUMBER OF PARTICLES (21.0 pm) /mlx104 O° 4181120116 2012426 100 120 Dust concentration (left) and 8(180) (right) variations in the D72 and D73 cores from Devon Island, N.W.T. The record in ice older than 10,000 years ts compiled by a combination of the two records. There ts probably a 5% loss of record due to tce flow effects in the passage of tee from 1 km uphill. 203 than present from 65,000 to 120,000 B.P. (Aharon et al. 1980; Shackleton and Opdyke 1973), which in turn, suggests the presence of larger ice volumes then than now. We have associated our warmer (less negative) 6 then as due to there being a different moisture source than the present one (Paterson et al. 1977). Furthermore, we think that the period before 70,000 B.P. in Figure 2 represents the last (Sangamon) interglacial. However, it is often considered that the last glaciation began some millenia before 70,000 B.P. when, indeed, sea levels were lower. This has been considered by Ruddiman et al. (1980) as possibly a warm period with high snow accumulation rates in the Arctic. This would place the beginning of the last glaciation at a part of our profile where there is no inflexion or major gradient change. The Ô data and the concept of a 120,000 B.P. start to the ice age are at present irreconcilable with the time scale depicted in Figure 2. Hollin (1980) argues that although sea levels were lower than present since 120,000 B.P. the main cooling stage and hence the major inception of the Laurentide ice sheet (well to the south of our drill site) was, in fact, about 70,000 B.P. He feels that the major but short inflection in Ô in our core (and Camp Century) at about 90,000 B.P. was due to an Antarctic surge which caused a short-lived but major global cooling due to an associated increase in global albedo.* There is much disagreement over the various proxy temperature records for these periods. Because of the highly unreliable time scale of our cores at these depths (a few metres above the bed of the ice cap), there is unfortunately very little that we can do with our D72 and D73 data to resolve the dilemma. However, we can pose an alternative explanation based on preliminary pollen data and a preliminary analysis of the Ag79 core. This is: High Arctic ice sheets began their growth about 115,000 B.P. when deep-sea sediment cores are interpreted to show a dramatic increase in global ice volume from interglacial (i.e. less than present-day) levels (Ruddiman et al. 1980). Sea levels are also believed to have begun dropping then (Aharon et al. 1980). Ice formed in the Canadian High Arctic would cause only a small percentage of the sea level change demanded by the ocean core evidence. Some support for this interpretation comes from the lack of pollen in all ice below the 10,000 B.P. 6 step (S. Lichti-Federovich, personal communication, 1977). Pollen begins to * Reflection of incoming radiation from the sun. 204 appear in increasing quantities only in the Holocene section (Lichti-Federovich 1977). Additionally, our most recent core (Ag79) has a 3-m basal layer of ice with substantial clay inclusions. This layer is unique to all our cores. The 3-m layer may have been formed during a glaciation of unknown age, but certainly one older than the last interglacial. The ice then had to be thick enough to cause basal melt at the Ag79 drill site. Thus, we visualize very limited ice forming the High Arctic ice caps during the last interglacial. For example, only the very highest parts of the Agassiz area would have born ice (e.g. the Ag79 drill site), whereas ground would have been clear of ice only 1 km down slope (e.g. at the Ag77 site). If we accept this view, it would be more logical to have ice caps beginning growth again at the initial stages of glaciation rather than during an interglacial, as Paterson et al. (1977) have indicated. In this case, early growth would have occurred under less negative 6 (i.e. relatively warm) conditions. In this case too, accumulation rates would have been substantially higher than present day ones because of the warmer conditions and higher availability of moisture. Further pollen analysis and study of the Ag79 core may help to resolve the problem. But a more likely source for a solution should be in the deep core at DYE 3 in Greenland that is presently being drilled by scientists from the United States and Denmark. However, the Devon cores can yield, and have yielded, valuable 6 (temperature) information over the past 10,000 years (Figure 3). We conclude from this profile that climate at the drill site has cooled by 2.7 to 3.5°C since 5,000 B.P. (Fisher and Koerner, in press). This value takes due consideration of the changing source and elevation effects on 6 referred to earlier. Other proxy temperature records are in general agreement with the 6 profile, such as those of Chu K'o-Chen (Newell 1974) for China, a Minnesota pollen record (Webb and Bryson 1972),a Mediterranean pollen core (Van der Hammen et al. 1971), and a North Atlantic plankton record (Sancetta et al. 1973). But a North Atlantic core (National Academy of Sciences 1975) and all the Antarctic ice cores (Barkov et al. 1975, Lorius et al. 1979, Epstein et al. 1970) show different trends. While North Atlantic core data are irreconcilable with our 5 profile, the difference between ours and the Antarctic 5 profile suggests that D72 and D73 Ô records are characteristic of the Northern Hemisphere and possibly that Antarctica has had a globally unique temperature trend over the past 10,000 years. 205 O -25 8 neo O -27 co ~~ -28 Le) 1000 2000 3000 4000 5000 10000 à Devon O À _29 2 | NO 4 -30 oO 0 J -2 O FIGURE 3: 206 Summer Temperature Northern Canada 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 YEARS BEFORE PRESENT The upper histogram represents the last 10,000-year record of 6(180) from Devon Island Ice Cap cores. The lower histogram represents summer temperatures of the last 7,500 years in northern Canada as indicated by pollen data (Nichols 1972). It is difficult to reconcile the Devon 6 profile with the Meighen Ice Cap analysis shown in Table 1, or to the 1961-80 mass balance of the northwest side of the Devon Island Ice Cap (Table 2). If there is a cooling trend for 5,000 years, it is difficult to place a warm period in the middle of it capable of giving negative balance conditions on Meighen Island 2,000 to 1,000 years ago. Moreover, the mass balance on the west side of the Devon Island Ice Cap is negative today, yet the D72 and 736 profiles indicate that it has generally been much warmer over all but about 5% of the last 5,000 years. If the balance is negative now, and yet the past 5,000 years have been generally warmer than today, this means that the Devon Island Ice Cap should have had a negative balance almost continuously for those 5,000 years. This is not the case, as not only our work around the margins of the ice cap (Fisher and Koerner, in press), but also that of Hattersley-Smith (1972) on northern Ellesmere Island, and Muller (1966) on Axel Heiberg Island show that these large ice caps are now close to their most advanced positions during the last 5,000 years. The answer may well lie in a poor relationship between 6 and mass balance. This could be because our 6 profile represents only the annual record. By studying the percentage of ice layers originating from summer melting near the Devon Island Ice Cap surface, it has been possible to determine variations in summer climate over the last 700 years (Koerner 1977). Results are shown in Figure 4. While the unique warmth at the first half of this century is clear from this record, so is the period of the Little Ice Age 150-300 years ago. The relationship between the melt and § records is not good. Firstly, there is an approximately 20- to 30-year lag between 6 trends and percent melt trends. Secondly, there is no sign of the earlier (i.e. down-core) warming trend in summer temperatures. Perhaps this is partly due to the shortness of the melt record. However, a pollen record from Arctic Canada (Nichols 1972) agrees only in general form to our 6 profile. The pollen record (Figure 3) shows a reasonable relationship to our Meighen Ice Cap data. Also, it does not conflict with the negative balance of the northwest side of the Devon Island Ice Cap today and the concept of a generally zero to slightly positive balance over most of the previous few thousand years. The pollen record shows a warmer period than present 1,000-2,000 years ago when Meighen Ice Cap was considered to have a negative balance, but a colder period 200-1,000 years ago when we infer a positive balance for Meighen. Admittedly, this is just one proxy temperature profile, but it suggests that the poorness of the 4/mass balance relationship lies in a more basic but weak relationship between the 207 YEARS B.P 200 400 600 800 Q < Fe D a * mA -28 + LL LL ll i i" mL b.) " AN TS ; = FIGURE 4: 6(180), percent melt (PC), and electrolytic conductivity (o) over the last 700-800 years from Devon Island Ice Cap cores. The blackened portions represent above-average values. 208 annual and the summer climate. It warrants further investigation probably through the longer summer melt records of Agassiz 77 and 79 cores. Superimposed on the 5,000-year 6 trend is a 2,500-year variation of § (see the residuals from the trend (Figure 5)). The 2,500-year trend has been related to a solar cycle, as there is a similar cycle in 14C in tree rings (Suess 1970). Production of 14C in the upper atmosphere is restricted when the solar wind shields the earth from cosmic rays. Solar wind strength is related to sun spot and solar flare activity. So, the link seems to be: increased sun spots and flares, increased solar constant (to a maximum at a critical sun spot number), decreased 14c and increased global temperatures. The relationship is shown in Figure 5. The Little Ice Age (about 150-300 years ago) coincides with the last residual trough (i.e. most negative 6, or colder temperatures). While some (e.g. Robock 1979) argue that the Little Ice Age relates most closely to increased volcanic activity, the effect of a change in solar activity cannot be ignored. We (Koerner and Fisher 1979) have attempted to predict temperature change over the next 50 years. This is the result of Fisher's (1976) work using Ô data for, say, the 2,000-1,000 B.P. period to predict the measured 1,000-0 B.P. 6 profile. Once the closest prediction is made, the same 6/t relationship is used to predict the future in terms of 6 and then temperature. Fisher's (1976) preliminary work suggests a continuation of the present cooling trend (in annual terms) to 1990 and perhaps beyond. The cooling is approximately 0.5-1.0°C. The same approach, using the core-melt percent data, predicts a summer cooling of some 1-2°C, bottoming out 20-50 years after the annual one. While numerical calculations are rather complex, the data indicate that summer conditions of such cold years as 1972 and 1976 will become more common in the High Arctic in future. The probable result will be shorter snow-free seasons and less navigable sea-ice conditions in the Queen Elizabeth Islands. The work of Hibler and Langway (1977) and Dansgaard et al. (1971) in Greenland are in general agreement with our predictions. Alt (1978), from a study of 14 years of synoptic weather charts between 1961 and 1974, has related weather systems to the Devon Island Ice Cap's mass balance. She finds that dominance of low pressure conditions over Baffin Bay not only reduces summer melting, but brings summer accumulation to the ice cap. Conversely, very negative balances result from 209 b ) c ) YEARS B.P O 1000 2000 3000 4000 5000 6000 7000 8000 | | Eddy’s a) 0 AC schematic Devon à residuals 0 Camp Century FIGURE 5: a) Persistent l*C deviations plotted schematically; + refers to increasing relative l*C and impltes decreasing solar activity (Eddy 1977). b) Deviattons of values from the 4,000-year trend line for the Devon Island 8(180) record. Beyond 5,300 B.P. the time scale is based on cross- correlation with the Camp Century core which ts accurate to within 5%. e) Deviations of values from the 8,000-year trend line for the Camp Century 8(180) record. 210 the development and persistence of anticyclonic blocking* and warm air advection.** Alt is presently attempting to view paleoclimate (as deduced from our ice cores) in terms of changes in the atmospheric circulation in the past. Chemistry and Dust The most dramatic variation in our cores, in terms of chemistry and dust, occurs prior to the Holocene (Figure 2). What we identify as the last glaciation in our core, and what should perhaps be referred to as the most negative § section (Figure 2), contains the highest dust and elemental concentrations (here we refer to Na, Mg, Ca, Al, Si and K). The dust particles are skewed to the smaller size range in the negative 6 section, which means that greater than normal amounts of dust and metals had a more distant source. The more extensive continental shelves that were exposed during the last glaciation have been considered as major sources (Fisher and Koerner, in press), as Ca (being derived from marine sediments) shows the greatest enrichment. The shelves reached a maximum exposure when sea level dropped to its Wisconsin minimum about 18,000 years ago. This is precisely when our cores show the highest concentration of dust and metals. Over the last 10,000 years we find no substantial trend or fluctuation in microparticle concentrations in D72 and D73. The concentration of Al in the Camp Century core also shows a flat profile for the Holocene. This means we cannot account for our 6 (temperature) trend since 5,000 B.P. in terms of increased or decreased turbidity in the atmosphere. We find increasing levels of Na and Mg into the past (Figure 6), but this is an effect rather than a cause of the 5 trend. It is reasonable to suppose that the effect on the drill site chemistry of not only the nearby North Watert to the east in Baffin Bay, but also the surrounding sea water, would be augmented with increasing warmth back in time. This is due to an increasing period of open water each year with warmer temperatures. In addition to Ed. Notes: * Presence of a ridge of high pressure over a region (see Lamb 1972, pp. 107-109). ** Transfer of heat by horizontal movement of air. + A large recurring polynia, or body of water largely surrounded by sea ice, located between southern Ellesmere Island and Greenland (see Dunbar and Dunbar 1972). 21 Devon D À OO @W Na b) ( g/kg) NO Ô FA d) conductivity (Sm°\) N oO je) EE il O0 1 2 3 + 5 6 THOUSANDS OF YEARS B.P FIGURE 6: Chemistry (Mg, Na and Ca) and electrolytic conductivity (0) from melt- tank water produced during drilling of D73 core on Devon Island, N.W.T. introducing more aerosols of marine origin into the local atmosphere, it would give rise to increased cloudiness, which we have found is often associated with higher levels of some metals (e.g. Na; see also Warburton and Linkletter 1977). Volcanic Activity Despite claims to the contrary (Thompson 1977) the Devon cores do not show a simple increase in micro-particle concentrations with known or postulated volcanic events. Hammer et al (1978, and in press) have found this to be true for Greenland cores also. So, volcanic ash of greater than 0.5 um (0.35 1m in Hammer et al.'s case) cannot be detected in Greenland or Canadian ice caps. Hammer et al. (1978) made a major advance in ice-core studies upon discovering that volcanic activity is detectable in cores through acid and not ash layers. Acid is deposited as acid snow following volcanic eruptions. At first Hammer used electrolytic conductivity measurements to pick out the acid layers. He then developed a new technique for detecting high Ht concentrations (i.e. low pH) in cores (Hammer et al., 1980). The method consists of rapidly scratching a clean ice-core surface with two electrodes fixed a few centimetres apart from each other and using a high voltage (1,000 volts). Acidity of the layer is roughly proportional to the current developed through it. Although we have a preliminary volcanic record using this technique for the Agassiz 77 core, we will refer here to our 5,000-year pH profile based on measurements of the melt-tank water from D73 generated by the thermal drill as it penetrates the ice. The profile shows no significant trend with very stable pH values averaging 5.46. Because the resolution of these measurements is no better than 5 years in more recently deposited ice and 50 years in the older ice, individual volcanic events cannot be picked out. However, the lack of a trend suggests that volcanic activity in terms of a 5-50 year resolution has not altered significantly over the last 5,000 years. Therefore our 6 (temperature) trend is not determined by changing levels of volcanic activity. The consensus of scientific opinion is that volcanic activity does cause atmospheric cooling, which is registered as about a 1-2°C cooling at the earth's surface lasting a few years after a major eruption. Hammer et al. (1980) have developed a 10,000-year record of volcanic activity as deduced from Greenland ice cores. They find a significant relationship between the level of background volcanic acid in Greenland cores and a Northern Hemisphere proxy-temperature profile covering the same period. Bradley and England (1978) attributed an apparent cooling trend in the High Arctic, beginning with a dramatic step in 1963, to the 213 eruption of Gonung Agung* in March, 1963. We have related individual events to temperature fluctuations and levels of melting in the core, but our results are preliminary. So far, from five major acid layers in the Agassiz 77 core we have found no evidence of an associated cooling (i.e. more negative 6) associated with these layers. Furthermore, we find no significant increase in acidity attributable to the Agung event in our D72 and D73 cores or Agassiz 77 and 79 cores. As we have not yet been able to relate any significant § change with our five major acid layers in the Agassiz 77 core, it seems unjustified to relate any post-1963 cooling to the Agung eruption, especially since we find several similar warm-to- cool jumps in the 1940s and 1950s in our summer melt percent record (Fisher and Koerner, in press). CONCLUSIONS The main conclusion we can draw from the D72 and D73 cores is that, generally, High Arctic climate has been cooling over the past 5,000 years. However, the 6 record represents annual temperatures, and a shorter record of summer melting in the same cores suggests that while the summer temperature changes lag behind the annual ones there may be a more complex and generally poor relationship between the two. Thus, the present negative balance of the Devon Island Ice Cap and the Meighen Ice Cap climatic record can be much better related to the summer melt record and, in particular, summer paleotemperatures as seen in certain Arctic pollen records. As far as summer climate is concerned, our cores show a very warm period 20-70 years B.P., and another from about 1,000-2,000 B.P. We are unable to determine the driving force behind the overall temperature decline at the drill sites, a trend which is in general agreement with other proxy temperature records elsewhere in the Northern Hemisphere. We have good evidence that neither increasing volcanic activity nor changing turbidity are the causes. We predict that the cooling in the High Arctic will continue for about another 50 years. The cooling will be of the order of * Ed. Note: For a general discussion of vulcanism in relation to glaciation during the last 40,000 years, see Bray (1974). 214 0.5-1°C. We think that summers will cool for about 20-50 years longer than the annual temperatures, and by about twice the amount of annual cooling. These predictions do not include any anthropogenic effects which may, for example in the case of co2 generation, counteract the cooling effect. Our future concerns are with our Ag77 and Ag79 cores where we have an improved time scale based partly on volcanic layers. Generally, our cores indicate that the large ice caps on Devon, Ellesmere and Axel Heiberg Islands are about 100,000-years, or more, old. The Meighen Ice Cap and Ward-Hunt Ice Shelf probably began their growth as conditions grew more favourable for glacierization 4,000 years ago. The remaining small ice caps are probably younger still, beginning growth at the end of the warm period 2,000-1,000 years ago which formed dirt layers on the Ward Hunt Ice Shelf and a relic ablation surface* on the Meighen Ice Cap. This is because a large number of the small ice caps are 40-m, or less, thick and probably would not have survived the 1,000-2,000 B.P. warm period. * Ed. Note: The remains of a former glacier surface where ice has been removed by melting. 215 REFERENCES Aharon, P., J. Chappell, and W. Compston. 1980. Stable isotopes and sea level data from New Guinea supports Antarctica ice-surge theory of ice ages. Nature 283 (5758): 649-651. Alt, B. 1978. Synoptic climate controls of mass-balance variations on Devon Island Ice Cap. Arctic and Alpine Res. 10(1): 61-80. - 1979. Investigation of summer synoptic climate controls on the mass balance of Meighen Ice Cap. Atmosphere-Ocean 17(3): 181-199. Barkov, N.I., F.G. Gordienko, Ye.S. Korotkevich, and V.M. Kotlyakov. 1975. Oxygen-isotope studies of the 500-meter ice core from the Vostok borehole. Inf. Bull. Soviet Antarctic Exped. 90: 34-49. Boutron, C., and C. Lorius. 1979. Trace metals in Antarctic snows since 1914. Nature 277 (5697): 551-554. Bradley, R.S., and J. England. 1978. Volcanic dust: influence on glacier mass balance at high latitudes. Nature 271: 736-738. Bray, J.R. 1974. Volcanism and glaciation during the past 40 millenia. Nature 252: 679-680. Dansgaard, We, S.J. Johnsen, H.B. Clausen, and C.C. Langway. 1971. Climatic record revealed by Camp Century ice core. In: The Late Cenozoic Glacial Ages. Edited by: K.K. Turekian. Yale Univ. Press, New Haven. pp. 37-56. Dunbar, M., and M.J. Dunbar. 1972. The history of the North Water. Proc. Roy. Soc. Edinburgh (B)72: 231-241. Eddy, J.A. 1977. Climate and the changing sun. Climatic Change 1(2): 173-190. Epstein, S., R.P. Sharp, and A.J. Gow. 1970. Antarctic ice sheet: stable isotope analyses of Byrd Station cores and interhemispheric climatic implications. Science 168: 1570-1572. Fisher, D.A. 1976. A study of two 6 (189) records from Devon Ice Cap, Canada, and comparison of them to Camp Century 6 record, Greenland. Ph.D. thesis, Univ. Copenhagen. 287 pp. Fisher, D.A., and ReM. Koerner. 1979. Some aspects of climatic change in the High Arctic during the Holocene as deduced from ice cores. Abstracts. Quaternary Climatic Change Symposium. York Univ., Toronto, 1979. p. 33. (Manuscript submitted for publication in Geo Abstracts volume "Quaternary Paleoclimate", edited by W.C. Mahaney). Hammer, C.U., H.B. Clausen, W. Dansgaard, N. Gundestrup, S.J. Johnsen, and N. Reeh. 1978. Dating of Greenland ice cores by flow models, isotopes, volcanic debris and continental dust. J. Glaciol. 20 (82): 3-26. Hammer, C.U., H.B. Clausen, and W. Dansgaard. 1980. Greenland ice sheet evidence of post-glacial volcanism and its climate impact. Nature 288 (5788): 230-235. Hattersley-Smith, G. 1972. Climatic change and related problems in Northern Ellesmere Island, N.W.T., Canada. In: Climatic Changes in Arctic Areas during the Last Ten-thousand Years. Edited by: Y. Vasari, H. Hyvärinen and S. Hicks. Acta Univ. Ouluensis, Ser. A, 3, Geol. 1, pp. 137-148. Hollin, J. 1980. Climate and sea level on isotope stage 5: an East Antarctic ice surge at 75,000 B.P.? Nature 283: 629-634. Hibler, W.D., and C.C. Langway. 1977. Ice core stratigraphy as a climatic indicator. In: Polar Oceans. Edited by: M.J. Dunbar. Arctic Inst. North Amer. Montreal. pp. 589-602. 216 Junge, C.E. 1963. Air chemistry and radioactivity. Academic Press, New York. 382 pp. - 1977. Processes responsible for the trace content in precipitations. In: Isotopes and Impurities in Snow and Ice. Proc. Grenoble Symposium, 1975, IASH Publ. 118: 63-77. Koerner, R.M. 1968. Fabric analysis of a core from the Meighen Ice Cap, Northwest Territories, Canada. J. Glaciol. 7: 421-430. - 1977. Devon Island ice cap; core stratigraphy and paleoclimate. Science 196: 15-18. Koerner, R.M., and D.A. Fisher. 1979. Liquified natural gas tanker transport through Parry Channel and climatic change. Internal Rept. Energy, Mines and Resources, Ottawa. 16 pp. Koerner, R.M., and W.S.B. Paterson. 1974. Analysis of a core through the Meighen Ice Cap, Arctic Canada, and its paleoclimatic implications. Quaternary Res. 4: 253-263. Koerner, R.M., and R.P. Russell. 1979. §189 variations in snow on the Devon Island Ice Cap, Northwest Territories, Canada. Can. J. Earth Sci. 16(7): 1419-1427. Lamb, H.-H. 1972. Climate: present, past and future. Vol. 1. Fundamentals and climate now. Methuen & Co. Ltd., London. 613 pp. Lichti-Federovich, S. 1975. Pollen analysis of surface snow from the Devon Island Ice Cap. Geol. Surv. Can. Paper 75-1A: 441-444. Lorius, Ce, Le Merlivat, J. Jouzel, and M. Pourchet. 1979. A 30,000-year isotope climatic record from Antarctic ice. Nature 280: 644-648. Lyons, J.B., and J.E. Mielke. 1973. Holocene history of a portion of northernmost Ellesmere Island. Arctic 26: 314-323. Muller, F. 1966. Evidence of climatic fluctuations on Axel Heiberg Island, Canadian Arctic Archipelago. In: Proceedings of the Symposium on the Arctic Heat Budget and Atmospheric Circulation. Edited by: J.0. Fletcher. Rand Corp., Santa Monica. pp. 136-156. National Academy of Sciences. 1975. Understanding climatic change. U.S. Committee for the Global Atmospheric Research Program, Nat. Res. Council, Washington, D.C. 239 pp. Newell, R.E. 1974. The earth's climatic history. Technol. Rev. 77(2): 2-17. Nichols, H. 1972. Summary of the palynological evidence of late Quaternary vegetational and climatic change in the Central and Eastern Canadian Arctic. In: Climatic Changes in Arctic Areas during the Last Ten-thousand Years. Edited by: Y. Vasari, H. Hyvärinen and S. Hicks. Acta Univ. Ouluensis, Ser. A, 3, Geol. 1, pp. 309-339. Paterson, W.S.B. 1968. A temperature profile through the Meighen Ice Cap, Arctic Canada. Internat. Assoc. Sci. Hydrol. Publ. 79: 440-449. + 1976. Vertical strain-rate measurements in an arctic ice cap and deductions from them. J. Glaciol. 17(75): 3-12. + 1980. Ice Sheets and ice shelves. In: Dynamics of Snow and Ice Masses. Edited by: S. Colbeck. Academic Press, New York. pp. 2-78. Paterson, W.S.B., R.M. Koerner, D. Fisher, S.I. Johnsen, H.B. Clausen, W. Dansgaard, P. Bucher, and M. Oeschger. 1977. An oxygen-isotope climatic record from the Devon Island Ice Cap, Arctic Canada. Nature 266(5602): 508-511. Robock, A. 1979. The "Little Ice Age": Northern Hemisphere average observations and model calculations. Science 206: 1402-1404. Ruddiman, W.F., A. McIntyre, V. Niebler-Hunt, and J.T. Durazzi. 1980. Oceanic evidence for the mechanism of rapid Northern Hemisphere glaciation. Quaternary Res. 13: 33-64. 217 Sancetta, C.J., J. Imbrie, and N.G. Kipp. 1973. Climatic record of the past 130,000 years in North Atlantic deep-sea core V23-82: correlation with the terrestrial record. Quaternary Res. 3: 110-116. Shackleton, N.J., and N.D. Opdyke. 1973. Oxygen isotope and paleomagnetic stratigraphy of Equatorial Pacific core V28-238: oxygen isotope temperatures and ice volumes on a 105 and 106 year scale. Quaternary Res. 3: 39-55. Suess, H.E. 1970. The three causes of secular cl4 fluctuations, their amplitudes and time constraints. In: Radio-carbon Variations and Absolute Chronology. Edited by: I.U. Olsson. Nobel Symposium 12. Wiley, New York. pp. 595-604. Thompson, L.G. 1977. Variations in microparticle concentration, size distribution and elemental composition found in Camp Century, Greenland, and Byrd Station, Antarctica, deep ice cores. In: Isotopes and Impurities in Snow and Ice. Proc. Grenoble Symposium, 1975, IASH Publ. 118: 351-364. Van der Hammen, T., T.A. Wijmstra, and W.H. Zagwijn. 1971. The floral record of the late Cenozoic of Europe. In: The Late Cenozoic Glacial Ages. Edited by: K. Turekian. Yale Univ. Press, New Haven. pp. 391-424. Warburton, J.A., and G.O. Linkletter. 1977. Precipitation-forming mechanisms and the chemistry of precipitation on the Ross Ice Shelf, Antarctica. In: Isotopes and Impurities in Snow and Ice. Proc. Grenoble Symposium, 1975. IASH Publ. 118: 88-94. Webb, T., and R.A. Bryson. 1972. Late- and post-glacial change in the northern midwest, USA: quantitative estimates derived from fossil pollen spectra by multivariate statistical analysis. Quaternary Res. 2: 70-111. 218 GORDON MANLEY MSc, MA 1902-1980 With the death of Gordon Manley on 29 January, the Society lost a Fellow of 53 years! standing, and British geography a scholar of individuality and world distinction. Born in Blackburn in 1902, Manley remained throughout his life very much a Lancastrian, loving the north country and deeply versed in its history and legends. From Queen Elizabeth's School in his native town, he proceeded to the University of Manchester to study engineering, and thence to Cambridge where he changed course to geography under Debenham. His growing interest in geography was further stimulated and focused by two early experiences after graduation: the first a short period of service in the Meteorological Office; the second, participation in an expedition to Greenland with James Wordie. Back in England he returned to the academic world, securing an Assistant Lectureship at Birmingham University; but shortly after, he migrated north to Durham as Lecturer and Head of its then tiny geography department. It was in Durham that he began the studies of British climate which remained a leading interest all his life, and for a lengthy period while there, he maintained virtually single-handed a meteorological station at Moorhouse, 545 m up on the bleak flanks of Cross Fell, accumulating the data which produced revealing new studies of the local helm wind. The historical studies of British climatic records, with which his name is so closely associated, were also begun in his Durham days. In 1939, however, he returned to Cambridge for nearly a decade, during which period he expanded his interests considerably in the realms of Pleistocene climates and glaciology, with fieldwork trips to Norway and elsewhere. In 1948, he moved on to Bedford College, London, as its first Professor of Geography, and had the experience for a second time of building up a smallish department, though now in the more favourable climate of general university expansion. With retiring age approaching, his unflagging energy led him to yet another move, to the Headship of the Department of Environmental Sciences at the new University of Lancaster. Retiring in 1968, he returned to Cambridge, but remained a Research Associate of Lancaster and actively engaged on research and writing until his death. By early training and natural inclination, Manley was in the tradition of field scientists and, although widely read and interested in literature and history, his geographical interests were directed strongly to the natural earth and its phenomena. Human 219 geography for him lacked sufficient rigorous discipline, and he was at times forthright in his criticisms. He consistently advocated first-hand investigation, on sound scientific principles, but in his chosen field of climatology after his early work on the Pennines, he was forced more and more to work with the observations of others — particularly as his interests moved into the field that he made his own: the collection, analysis and interpretation of early instrumental climatic records. He was indefatigable in searching these out, and, with their aid, he made many important contributions to the history of climate in Britain. But he had equal interests in the wider scene, and was a frequent contributor at international conferences on global climatic history. Among his many scientific publications, his book Climate and the British Scene (New Naturalist Library, 1952) is probably the best known to the wider public. In later years he appeared frequently on radio and on television in discussions on meteorological events and climatic trends. Manley's lifelong and enthusiastic studies in the fields of meteorology and climate raised him to the rank of an acknowledged international authority, and were recognized by a number of awards, including the Leverhulme Award in 1937, the Buchan Prize of the Royal Meteorological Society in 1943, and the Murchison Award of the Royal Geographical Society in 1947. He was correspondent for glaciology to the British National Committee for the International Geophysical Year (1955-61) and, in the period 1964-69, Visitor to the Ministry of Technology. The distinction he particularly prized, however, was his election to the Presidency of the Royal Meteorological Society in 1945-46, and he remained an active and productive Fellow of that Society up to his death. Widely known and respected in the geographical and meteorological worlds, Manley was an effervescent and entertaining conversationalist, and a lively and stimulating companion. His appetite for converse was never-failing; for years after the event, those present spoke with awe of his achievement in talking non-stop throughout the entire ascent of the 2000-odd metres of Galdhoppigen in the Jotunheim. He married in 1930 Audrey Fairfax, daughter of the late Professor Arthur Robinson, sometime Master of Hatfield College, University of Durham, who survives him. His departure removes a notable figure from the ranks of a subject currently enjoying a marked revival of interest, a revival which he did much to foster. (With acknowledgements to The Geographical Journal) 220 RECENT SYLLOGEUS TITLES / TITRES RECENTS DANS LA COLLECTION SYLLOGEUS No. 21° Brunton, D:F-1(1979) THE VASCULAR PLANT COLLECTIONS OF JOHN MACOUN IN ALGONQUIN PROVINCIAL PARK, ONTARIO. 20 p. No. 22 Warkentin, John (1979) GEOLOGICAL LECTURES OF DR. JOHN RICHARDSON, 1825-26. 63 p. No. 23 Cody, William J. (1979) VASCULAR PLANTS OF RESTRICTED RANGE IN THE CONTINENTAL NORTHWEST TERRITORIES, CANADA. SA Ds No. 24 Haber, Erich and James H. Soper (1979) VASCULAR PLANTS OF GLACIER NATIONAL PARK, BRITISH COLUMBIA, CANADA. No. 25 Danks, H.V. (1980) ARTHROPODS OF POLAR BEAR PASS, BATHURST ISLAND, ARCTIC CANADA. ii, 68 p. No. 26 Harington, C.R. (ed.) (1980) CLIMATIC CHANGE IN CANADA. 246 p. No. 27 White, David J. and Karen L. Johnson (1980) THE RARE VASCULAR PLANTS OF MANITOBA. / LES PLANTES VASCULAIRES RARES DU MANITOBA. 52 p. i / 53: ps No. 28 Douglas, George W., George W. Argus, H. Loney Dickson, & Daniel F. Brunton (1981) THE RARE VASCULAR PLANTS OF THE YUKON. / LES PLANTES VASCULAIRES RARES DU YUKON. No. 29 Brodo, I.M. (1981) LICHENS OF THE OTTAWA REGION. (English edition) LICHENS DE LA REGION D'OTTAWA. (L'édition frangaise) No. 30 Manning, T.H. (1981) BIRDS OF TWIN ISLANDS, JAMES BAY, N.W.T., CANADA. No. 31 Ferguson, R.S. (1981) SUMMER BIRDS OF THE NORTHWEST ANGLE PROVINCIAL FOREST AND ADJACENT SOUTHEASTERN MANITOBA. No. 32 Ouellet, H. et F.R. Cook (1981) LES NOMS FRANCAIS DES AMPHIBIENS ET REPTILES DU CANADA: UNE LISTE PROVISOIRE. hes ” | | | an *) Mae CALIF ACAD OF SCIENCES LIBR | i 3 1853 10004 4663 | ui We